[racket] syntax-parse #:at

I don't understand your objections to the `printf' results, but I think
I can now better explain your original problem.

When the template has

   (syntax-parse stx
    [(_ new-name)
     (with-syntax ([output (make-transformer name pattern template)])
       #'(define-syntax new-name output))])

then the expansion of

  (define-new-syntax x ... ...)

has 

   (syntax-parse stx
    [(_ new-name)
     (with-syntax ([output (make-transformer x pattern template)])
       #'(define-syntax new-name output))])

There's an `x' as the fist sub-form to `make-transformer', as intended.


When the `new-name's are changed to `name' so that the template has

   (syntax-parse stx
    [(_ name)
     (with-syntax ([output (make-transformer name pattern template)])
       #'(define-syntax name output))])

then the expansion of

  (define-new-syntax x ... ...)

has 

   (syntax-parse stx
    [(_ x)
     (with-syntax ([output (make-transformer x pattern template)])
       #'(define-syntax x output))])

Again, `x' appears in as the first sub-form of `make-transformer' ---
but this `x' is bound as a pattern variable by the enclosing
`syntax-parse', which makes it a different `x' than the one when
`new-name' is used in the initial template.

At Mon, 21 May 2012 11:10:59 -0600, Jon Rafkind wrote:
> On 05/21/2012 10:06 AM, Jon Rafkind wrote:
> > On 05/21/2012 06:02 AM, Matthew Flatt wrote:
> >> At Sun, 20 May 2012 22:42:30 -0600, Jon Rafkind wrote:
> >>> The issue is why do I have to use a name other than 'name' for the
> >>> pattern variable on the line where it says 'HERE'. If I use the
> >>> 'name' pattern variable then the syntax-parse form generated by
> >>> 'make-transformer' will use the wrong lexical context for the literal
> >>> set. I don't see how the lexical context of the outer 'name' differs
> >>> from the inner 'new-name', other than the 'new-name' has an extra
> >>> mark on it.
> >> When you use `name' instead of `new-name', then `name' gets replaced by
> >> `x' from the use of `define-new-syntax', and `x' (unlike `name') has a
> >> different lexical context than `new-name'.
> >>
> >
> > I thought that too but while debugging it didn't seem to be the case. If I put a 
> printf in the 'make-transformer' macro to "see" what name is being passed in then 
> using 'name' or 'new-name' as the pattern variable results in the same thing being 
> printed.
> >
> >   (begin-for-syntax
> >     (define-syntax (make-transformer stx)
> >       (syntax-parse stx
> >         [(_ name pattern template)
> >          (printf "Name is ~a\n" #'name)
> >          #'#'(lambda (stx)
> >              (syntax-parse stx
> >                #:literal-sets ([literals #:at name])
> >                [pattern template]))])))
> >
> >
> > Then calling 'make-transformer' 3 different ways:
> >
> > (lambda (stx)
> >   (syntax-parse stx
> >    [(_ new-name)
> >     (with-syntax ([output (make-transformer name pattern template)])
> >       #'(define-syntax new-name output))]
> >
> > "Name is #<syntax:x.rkt:56:19 x>", where line 56 is the use-site of 
> 'define-new-syntax'
> >
> > (lambda (stx)
> >   (syntax-parse stx
> >    [(_ name)
> >     (with-syntax ([output (make-transformer name pattern template)])
> >       #'(define-syntax name output))]
> > "Name is #<syntax:x.rkt:56:19 x>"
> 
> 
> I guess this case doesn't print #<syntax:x.rkt:39 x> because the 'name' at the 
> use-site of 'make-transformer' is not bound by the pattern variable 'name'. If that 
> is the case then it seems this case should be identical to the 1st case where it 
> uses a different pattern variable, 'new-name'.
> ____________________
>   Racket Users list:
>   http://lists.racket-lang.org/users