On Sep 14, 2009, at 12:22 PM, Brian Harvey wrote:
> (Also of course, what's most important isn't that R7RS say "there
> must be
> a REPL," but rather that it refrain from specifying semantic
> constraints
> ("two-pass", e.g.) that are difficult or impossible to provide in a
> REPL.
> And let's not go back to the bad old days when compiled Lisp had
> different
> semantics (lexical scope) from interpreted Lisp (dynamic scope).)
Some R5RS Schemes provide two-pass semantics when forms entered at the
REPL are wrapped in a `begin':
(begin (define foo (bar)) (define-syntax bar (syntax-rules () ((_)
1))) foo)
-> 1 in Scheme48, MIT and Gauche. It's an error in PLT (R5RS mode),
Larceny (R5RS mode), Gambit (hygienic mode), Chez 7.4d, SISC, and
Bigloo.
I believe that this approach is sensible. It treats `begin' as a
binding form in all contexts, with unified semantics in each
situation. This only leaves open the question of how to treat `load'ed
files / R6RS-style top-level programs: as wrapped in an implicit
`begin', or a sequence of forms as if entered at the REPL?
--
Brian Mastenbrook
[email protected]
http://brian.mastenbrook.net/
_______________________________________________
r6rs-discuss mailing list
[email protected]
http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
