[racket-dev] can a syntax-class "inherit" attributes?

From: Stephen Chang (stchang at ccs.neu.edu)
Date: Fri Aug 23 12:53:30 EDT 2013

Ah, thanks. That's exactly what I wanted.

How stable are these features? Shall I add them to the docs?

On Fri, Aug 23, 2013 at 12:54 AM, Carl Eastlund <cce at ccs.neu.edu> wrote:
> Here's how you want to define AorB:
> (define-syntax-class AorB #:auto-nested-attributes
>   (pattern :A)
>   (pattern :B))
> There's two things going on here, both of which I believe are undocumented,
> sadly.
> One is #:auto-nested-attributes, an option that tells AorB that anything
> bound as an attribute by nested sub-patterns in its clauses should
> automatically become an attribute of AorB.
> The other is that empty identifier does not add "." to its attribute names.
> When you bind a with the attribute X, you get A.X, and when you bind b, you
> get b.X, but when you bind the empty identifier, you just get X.
> So now you just have X bound in both clauses, and that's automatically made
> an attribute of AorB, and it all works like you want.
> Carl Eastlund
> On Fri, Aug 23, 2013 at 12:22 AM, Stephen Chang <stchang at ccs.neu.edu> wrote:
>> Is there a way for a syntax-class to automatically inherit attributes?
>> For example, the third class below combines the first two. Is there a
>> way to automatically get the attributes from the first two classes in
>> the third one (ie, I want to drop the "#:attr X #'a.X" part in the
>> third class)?
>> #lang racket
>> (require syntax/parse)
>> (define-syntax-class A (pattern (a b) #:attr X #'b))
>> (define-syntax-class B (pattern (a b c) #:attr X #'c))
>> (define-syntax-class AorB
>>   (pattern a:A #:attr X #'a.X)
>>   (pattern b:B #:attr X #'b.X))
>> (syntax-parse #'(list 1) [ab:AorB (attribute ab.X)])
>> _________________________
>>   Racket Developers list:
>>   http://lists.racket-lang.org/dev

Posted on the dev mailing list.