[plt-scheme] Infix to prefic

I am going to look into this.  Have you tried it with plt-match?
Is there a difference?

Bruce



Bruce

On Sun, 23 May 2004, Jens Axel Søgaard wrote:

>   For list-related administrative tasks:
>   http://list.cs.brown.edu/mailman/listinfo/plt-scheme
>
> Anton's nice infix macros on comp.lang.scheme made
> me brush up some old code original used to have
> a convenient input syntax for polynomials.
>
> (module infix mzscheme
>    (require (lib "match.ss" "mzlib"))
>    (provide infix->prefix)
>
>    (define (infix->prefix exp)
>      ; Convert an infix expression to prefix
>      (define ip infix->prefix)
>      (define (nott sym)  (lambda (x) (not (eq? x sym))))
>      (define (non-op? s) (not (member s '(+ - * / ^))))
>      (match exp
>        [(? number? r)                          r]                     ; number
>        [(? symbol? s)                          s]                     ; variable
>        [((? non-op? s) (x ...))                `(,s ,@(map ip x))]    ; function call
>        [(p)                                    (ip p)]                ; parentheses
>        [('- exp)                               `(- ,(ip exp))]        ; sign
>        [('+ exp)                               `(+ ,(ip exp))]        ; sign
>        [((and p (? (nott '+))) ... '+ q ...)   `(+ ,(ip p) ,(ip q))]
>        [((and p (? (nott '-))) ... '- q ...)   `(- ,(ip p) ,(ip q))]
>        [((and p (? (nott '*))) ... '* q ...)   `(* ,(ip p) ,(ip q))]
>        [((and p (? (nott '/))) ... '/ q ...)   `(/ ,(ip p) ,(ip q))]
>        [((and p (? (nott '^))) ... '^ q ...)   `(^ ,(ip p) ,(ip q))]
>        [_  (error "infix->prefix: The following is not a legal infix
>                    representation of an arithmetical expression: " exp)]))
> )
>
>  >(require (lib "defmacros.ss"))
>  >(define-macro (: . infix)
>     (infix->prefix infix))
>  >(define x 3)
>  > (: sin(x) + 1 * tan(2))
> -2.0439198552016515
>
> The observant reader will by now have noticed that / and ^ associate
> the wrong way, but that's not my problem. My problem is the expansion
> time of the above match expression. If the two lines with / and ^ is
> commented out the above is evalueted almost immediately.
>
> Is there a more a suitable way to write the match expression?
>
> --
> Jens Axel Søgaard
>
>
>