Re: new function

[Damien Mattei <damien.mattei@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 5996 bytes --]

so i drop off the idea of a single assignment operator, there will be 2 ,<-
for single assignment when the variable has been previously defined (at
upper level), and another <+ that add a variable in the environment using
define.
Other possibility is to have a single assignment operator <- and to declare
the variables not previously defined at upper level.
All this was already done in Scheme+ :

here is what looks like
 the module definition :
(define-module (Scheme+Guile)

  #:export (def $bracket-apply$ <- -> <+ declare $ condx <> )

  )

(include "def.scm")
(include "Scheme+Guile/array-square-brackets.scm")
(include "Scheme+Guile/assignment.scm")
(include "declare.scm")
(include "condx.scm")
(include "block.scm")
(include "not-equal.scm")

and so, the code for the 2 assignment operator :

*Preview:*

(define (subset-sum-guile L t)

  {ls <+ (length L)}
  {dyn <+ dyna[ls t]}

  {cpt <- {cpt + 1}} ;; cpt has been already defined at toplevel

  ;; dyna[ls][t] means 0: unknown solution, 1: solution found, 2: no solution

  (condx [{dyn <> 0} (one? dyn)]
	 [(null? L) {dyna[ls t] <- 2}  #f] ;; return #f
	
	 [exec {c <+ (first L)}]	
	 ;; c is the solution
	 [{c = t} {dyna[ls t] <- 1}  #t]  ;; return #t
	
	 [exec {R <+ (rest L)}]	
	 ;; continue searching a solution in the rest
	 [{c > t} {s <+ (subset-sum-guile R t)}
	          {dyna[ls t] <- (one-two s)}
		  s] ;; return boolean value
			
	 ;; else : c < t at this point
	 ;; c is part of a solution OR not part of a solution
	 [else {s <+ {(subset-sum-guile R {t - c}) or (subset-sum-guile R t)}}
	       {dyna[ls t] <- (one-two s)}
	       s])) ;; return boolean value


and the code which use a single operator and a *declare* (it is a define
unspecified):

*Preview:*

(define (subset-sum-guile-dec L t)

  (declare ls dyn c R s)

  {ls <- (length L)}
  {dyn <- dyna[ls t]}

  {cpt <- {cpt + 1}} ;; cpt has been already defined at toplevel

  ;; dyna[ls][t] means 0: unknown solution, 1: solution found, 2: no solution

  (condx [{dyn <> 0} (one? dyn)]
	 [(null? L) {dyna[ls t] <- 2}  #f] ;; return #f
	
	 [exec {c <- (first L)}]	
	 ;; c is the solution
	 [{c = t} {dyna[ls t] <- 1}  #t]  ;; return #t
	
	 [exec {R <- (rest L)}]	
	 ;; continue searching a solution in the rest
	 [{c > t} {s <- (subset-sum-guile-dec R t)}
	          {dyna[ls t] <- (one-two s)}
		  s] ;; return boolean value
			
	 ;; else : c < t at this point
	 ;; c is part of a solution OR not part of a solution
	 [else {s <- {(subset-sum-guile-dec R {t - c}) or (subset-sum-guile-dec R t)}}
	       {dyna[ls t] <- (one-two s)}
	       s])) ;; return boolean value


the two syntax can be mixed but i do not think it is a good idea to have 2
coding style in the same project, also the position of 'declare' or '<+' is
depending of the possibility of the used Scheme implementation to have
define placed everywhere more or less, in the worse case the 'declare' or
'<+' have to placed at top of the body of the procedure.
The way python declare them, note that i have no regret to not be able to
define variables nested in a branch of code the way python do it that can
be seen from the whole procedure.I do not think it is a good idea and
useful. But i hoped to be able to avoid a 'declare' at the top or somewhere
but declaration even in Scheme could be necessary with some typed Scheme or
when using object oriented features of some Scheme perheaps that used typed
variables (for overloading function? not sure it exist in Scheme also ,even
in OOP).So this is not a great problem.

Damien


On Thu, Sep 23, 2021 at 11:53 PM Damien Mattei <damien.mattei@gmail.com>
wrote:

> yes i'm using GNU Guile 3.0.7,
> https://www.gnu.org/software/guile/docs/master/guile.html/Syntax-Transformer-Helpers.html
>
> i have tested a lot , even define-once again and i choose to use to
> assignment operators and portable code because the non-portable function do
> not bring more, finally it was not a bad idea to ask for a new function
> because we can do it with the actual toolbox...
> Damien
>
> On Thu, Sep 23, 2021 at 10:48 PM Taylan Kammer <taylan.kammer@gmail.com>
> wrote:
>
>> Responding to myself:
>>
>> On 23.09.2021 22:27, Taylan Kammer wrote:
>>
>> > I can't seem to find syntax-local-binding in Guile 2.2 or 3.0.  Did you
>> > have to import some special module, or are you using another version?
>>
>> Worked when I imported (system syntax internal).
>>
>> > Either way, I suspect that the following will not work with your macro:
>> >
>> >   (let ()
>> >     (let ()
>> >       (<$ x 1))
>> >     (display x)
>> >     (newline))
>>
>> Indeed it doesn't work, though for a different reason:
>>
>>   While compiling expression:
>>   Syntax error:
>>   unknown file:43:8: body should end with an expression in form (let ()
>> (<$ x 1))
>>
>> That's because indeed the inner let expands into:
>>
>>   (let ()
>>     (define x 1))
>>
>> And there has to be at least one expression after the define.  So I tried:
>>
>>   (let ()
>>     (let ()
>>       (<$ x 1)
>>       (newline))
>>     (display x)
>>     (newline))
>>
>> And as I expected, it says 'x' is unbound:
>>
>>   ;;; <stdin>:44:45: warning: possibly unbound variable `x'
>>   <$ : global scope : x
>>
>>   ice-9/boot-9.scm:1685:16: In procedure raise-exception:
>>   Unbound variable: x
>>
>> The only way it will work is if you never use nested scopes, but that will
>> lead to very strange Scheme code, and there will probably be many cases
>> where you accidentally use a nested scope without immediately noticing it.
>>
>> Note also that definitions aren't allowed everywhere.  Consider this:
>>
>>   (let ()
>>     (if 'whatever
>>         (<$ x 1)
>>         (<$ x 2))
>>     (display x)
>>     (newline))
>>
>> It leads to:
>>
>>   While compiling expression:
>>   Syntax error:
>>   unknown file:49:17: definition in expression context, where definitions
>> are not allowed, in form (define x 1)
>>
>> Because the arms of 'if' aren't allowed to be definitions.
>>
>> --
>> Taylan
>>
>

[-- Attachment #2: Type: text/html, Size: 24740 bytes --]