The reader tokenizes before it tries to make sense of a # sequence, so
#&( will be a missing-close-paren error long before you get to the
*#readers* handlers. For most other cases you might be able to use
port-position which is settable:
(set! *#readers*
(cons
(cons #\& (lambda (str)
(if (= (string-length str) 1) ; #& followed by a reader
delimiter
(read)
(begin
(set! (port-position (current-input-port))
(- (port-position (current-input-port))
(string-length str)
-1))
(string (string-ref str 1))))))
()))
(with-input-from-string "#&N"
(lambda ()
(display (read)) (newline)
(display (read)) (newline)))
N
#<eof>
(with-input-from-string "#&Newton"
(lambda ()
(display (read)) (newline)
(display (read)) (newline)
(display (read)) (newline)))
N
ewton
#<eof>
(with-input-from-string "#&N#&123"
(lambda ()
(display (read)) (newline)
(display (read)) (newline)
(display (read)) (newline)
(display (read)) (newline)))
N
1
23
#<eof>
(with-input-from-string "#&(123)"
(lambda ()
(display (read)) (newline)
(display (read)) (newline)))
(123)
#<eof>
I don't want the arity of the *#reader* function to be a flag meaning
"read one character" -- this would only have one use, would require
hand-waving in the documentation, and would make error checking less
reliable.
_______________________________________________
Cmdist mailing list
[email protected]
https://cm-mail.stanford.edu/mailman/listinfo/cmdist
