[Larceny-users] Probably not a bug in R6RS, but...

[Felix Klock]

David (cc'ing larceny-users)-

On Dec 15, 2008, at 5:57 AM, David Rush wrote:

> And yes, I didn't *really* mean to make #\space into a symbol that was
> equivalent to LAMBDA in the first place, but I still think that this
> half-functional situation is odd. It looks like maybe there is some
> interference with the lexer...

Yes, I concur that this is likely to be a lexing/parsing issue.

% ~/bin/larceny
Larceny v0.964+ "Fluoridation" (Aug 12 2008 22:00:42, precise:Posix  
Unix:unified)
larceny.heap, built on Tue Aug 12 22:01:04 EDT 2008

 > (define-syntax \ (syntax-rules () ((\ v e0 e+ ...) (lambda v e0 e 
+ ...))))
\x20;

 > (macro-expand '(\ v  (apply + v)))
((lambda () (\x20;v (apply + v))))

 > (macro-expand '(\  v (apply + v)))
((lambda ()
    (lambda \x2e;v\x7c;1\x7c;2 (begin v (apply + v)))))

Funny stuff.

Oh, actually, I think I just figured out your problem: the macro  
pattern (\ v e0 e+ ...) is treating the entirety of '\ v' as a single  
token.

So if you like, you can try this instead:

% ~/bin/larceny
Larceny v0.964+ "Fluoridation" (Aug 12 2008 22:00:42, precise:Posix  
Unix:unified)
larceny.heap, built on Tue Aug 12 22:01:04 EDT 2008

 > (define-syntax \ (syntax-rules () ((\  v e0 e+ ...) (lambda v e0 e 
+ ...))))
\x20;

 > (\  (a b) (+ a b))
#<PROCEDURE>

 > ((\  (a b) (+ a b)) 1 2)
3


Note that in the latter, I am carefully placing at least two spaces  
after the '\' character.

(As for your question of whether it should be filed as a bug -- I  
figure the worst that can happen is that we'll assign in a status like  
"wontfix" or "notabug".  I do ask though that if you do file it as a  
bug, make sure to describe what the expected behavior should be -- the  
only reason I have not filed a bug on this myself is that it is not  
immediately obvious what behavior one should expect here -- the  
current behavior may be entirely correct, and the problem is that your  
macro was incorrectly written, as I describe above.)

-Felix