> Well, that's the thing. +%bignums (metalanguage for "the 'bignums'
> feature group is present) means that exact integer overflow is an error,
> whereas -%bignums means exact integers overflow into inexact results.
I remember the discussion that led to this conclusion, but it remains
extremely counterintuitive to me, and I'm betting to most Scheme users.
"Bignums" means that (factorial 100) has an exact answer. There might
well be some even larger number that becomes inexact. By your definition
an implementation that errors at 2^31 "has bignums." Nah.
Anyway, this (what to do on overflow) might well be something the user
would like to /set/, as in
(set! auto-convert-to-float #f)
whereas the user can't turn on and off the capability of the software to
do more-than-32-bit-integer precision. /That's/ more of a %feature as
you would put it. :-)
_______________________________________________
r6rs-discuss mailing list
[email protected]
http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
