Re: Unexpected behavior

[Lindsay John Lawrence]

Thanks Alex!

This makes sense to me now...

(de Fibonacci (N)
   (let (F (list 2 1 1))
  (cond
 ((= N 0) 0)
 ((= N 1) (caddr F))
 ((= N 2) (cadr F))
 (T
(do (- N 3)
   (rot F)
   (set F (+ (cadr F) (caddr F))) )
(car F) ) ) ) )

/Lindsay


On Sat, Mar 11, 2017 at 11:35 AM, Alexander Burger 
wrote:

> On Sat, Mar 11, 2017 at 11:20:21AM -0800, Lindsay John Lawrence wrote:
> > Nm... After tinkering in the debugger.. using 'pp etc..
> > I get it. I am still getting the  (= code data) in picolisp.
>
> Cool! Our mails overlapped! :)
>
> ♪♫ Alex
> --
> UNSUBSCRIBE: mailto:picolisp@software-lab.de?subject=Unsubscribe
>

Re: Unexpected behavior

[Lindsay John Lawrence]

Or strictly speaking...

(de Fibonacci (N)
   (let (Fib '(2 1 1))
  (set Fib 2 (cdr Fib) 1 (cddr Fib) 1)
  (cond
 ((= N 0) 0)
 ((= N 1) (caddr Fib))
 ((= N 2) (cadr Fib))
 (T
(do (- N 3)
   (rot Fib)
   (set Fib (+ (cadr Fib) (caddr Fib)))
)
(car Fib))
 ) ) )

/Lindsay



On Sat, Mar 11, 2017 at 11:20 AM, Lindsay John Lawrence <
lawrence.lindsayj...@gmail.com> wrote:

> Nm... After tinkering in the debugger.. using 'pp etc..
> I get it. I am still getting the  (= code data) in picolisp.
>
> This gives the results I wanted.
>
> (de Fibonacci (N)
>(let (Fib '(1 1 0) R 0)
>   (do N
>  (rot Fib)
>  (set Fib (+ (cadr Fib) (caddr Fib)))
>   )
>   (setq R (car Fib))
>   (set Fib 1 (cdr Fib) 1 (cddr Fib) 0)
>   R
> ) )
>
> Now I can write...
>
> : (Fibonacci 1)
> -> 2
> : (Fibonacci 10)
> -> 144
> : (bench (nil (Fibonacci 10)))
> 3.388 sec
> : (length (chop (Fibonacci 10)))
> -> 20899 digits!
>
> Picolisp bignums are awesome!
>
> /Lindsay
>
>
>

Re: Unexpected behavior

[Alexander Burger]

Hi Lindsay,

> (de Fibonacci (N)
>(let (Fib '(1 1 0))
>   (do N
>  (rot Fib)
>  (set Fib (+ (cadr Fib) (caddr Fib))) )
>(car Fib) ) )
> 
> The results are something like a 'co routine or 'job... in that Fib as a
> memory between calls and it is not re-initialized with the 'let as I
> expected.

This is because 'rot' is a destructive operation. The list (1 1 0), i.e. these
three *cells*, are a constant in the body of the function (please 'F' -> 'f' ;)

You can see this if you do (pp 'fibonacci) after you executed it.

As in all such cases of destructive operations, the right way is to use a fresh
(newly built) list:

   (de fibonacci (N)
  (let Fib (list 1 1 0)  # builds a new list
 (do N
...

♪♫ Alex
-- 
UNSUBSCRIBE: mailto:picolisp@software-lab.de?subject=Unsubscribe

Re: Unexpected behavior

[Alexander Burger]

On Sat, Mar 11, 2017 at 11:20:21AM -0800, Lindsay John Lawrence wrote:
> Nm... After tinkering in the debugger.. using 'pp etc..
> I get it. I am still getting the  (= code data) in picolisp.

Cool! Our mails overlapped! :)

♪♫ Alex
-- 
UNSUBSCRIBE: mailto:picolisp@software-lab.de?subject=Unsubscribe

Re: Unexpected behavior

[Lindsay John Lawrence]

Nm... After tinkering in the debugger.. using 'pp etc..
I get it. I am still getting the  (= code data) in picolisp.

This gives the results I wanted.

(de Fibonacci (N)
   (let (Fib '(1 1 0) R 0)
  (do N
 (rot Fib)
 (set Fib (+ (cadr Fib) (caddr Fib)))
  )
  (setq R (car Fib))
  (set Fib 1 (cdr Fib) 1 (cddr Fib) 0)
  R
) )

Now I can write...

: (Fibonacci 1)
-> 2
: (Fibonacci 10)
-> 144
: (bench (nil (Fibonacci 10)))
3.388 sec
: (length (chop (Fibonacci 10)))
-> 20899 digits!

Picolisp bignums are awesome!

/Lindsay



On Sat, Mar 11, 2017 at 10:49 AM, Lindsay John Lawrence <
lawrence.lindsayj...@gmail.com> wrote:

> Hi
>
> I am having difficulty understanding the following..
>
> This works as expected...
>
> : (setq Sum '(1 1 0))
> -> (1 1 0)
> : (set Sum (+ (cadr Sum) (caddr Sum)))
> -> 1
> : (rot Sum)
> -> (0 1 1)
> : (set Sum (+ (cadr Sum) (caddr Sum)))
> -> 2
> : (rot Sum)
> -> (1 2 1)
> : (set Sum (+ (cadr Sum) (caddr Sum)))
> -> 3
> : (rot Sum)
> -> (1 3 2)
> : (set Sum (+ (cadr Sum) (caddr Sum)))
> -> 5
> : (rot Sum)
> -> (2 5 3)
> : (set Sum (+ (cadr Sum) (caddr Sum)))
> -> 8
> : Sum
> -> (8 5 3)
>
> However, when I put it in a function...
>
> (de Fibonacci (N)
>(let (Fib '(1 1 0))
>   (do N
>  (rot Fib)
>  (set Fib (+ (cadr Fib) (caddr Fib))) )
>(car Fib) ) )
>
> The results are something like a 'co routine or 'job... in that Fib as a
> memory between calls and it is not re-initialized with the 'let as I
> expected.
>
> -> Fibonacci
> : (Fibonacci 1)
> -> 2
> : (Fibonacci 1)
> -> 3
> : (Fibonacci 1)
> -> 5
> : (Fibonacci 1)
> -> 8
> ...
>
> The behavior I expected was
>
> : (Fibonacci 3)
> -> 5
> ..
>
> Why?
>
> /Lindsay
>
>
>
>
>