[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Distinct fields

Andre van Tonder <andre@xxxxxxxxxxxxxxxxx> writes:

> The issues section states:
> "Macros that expand into the implicit-naming layer might have unexpected 
> behavior, as field names that are distinct as identifiers may not be distinct 
> as symbols, which is how they're used."
> ^^^^^^^^^^
> Is the concept of distinctness defined with respect to symbolic equivalence?

Not so far as I know.

> It is not obvious to me that this should be the case.
> In the context of syntax-case, my first thought was that perhaps
> free-identifier=? might be a more appropriate predicate for
> distinctness, but then the following occurred to me:
> (define-syntax foo
>   (syntax-rules ()
>     ((_ f)
>      (define-record-type p
>        (fields (mutable x)       ; - would make a secret p-x
>                (mutable f))))))  ; - would make a toplevel p-x
>                                  ;   when called as below
> (foo x)
> In implementations of syntax-case that treat defines on
> macro-generated identifiers as establishing "secret" bindings
> (such as Chez and PLT), the p-x bindings for the two fields would
> be distinct, and not interfere, even though they are
> free-identifier=?  So the example would probably work.

That would depend on the way the macro defining the implicit
variant of DEFINE-RECORD-TYPE is written, right?  In particular, I
think the ones that I wrote decompose the names into strings and then
back, so I think any distinctness might get lost on the way.

Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla