Marcin 'Qrczak' Kowalczyk wrote:
Alan Watson <a.watson@xxxxxxxxxxxxxxxx> writes:I aplaud this goal. However, rather than create this horrid mess of type-specific operators, I would much prefer an approach that used type declarations, perhaps: (with-flonum (x y) (/ x y)) (with-fixnum (n m) (* n m))Declaring types of identifiers is not enough, because it can't express the wish to ignore any fixnum overflow and assume that the result will fit in a fixnum.
It can - if we define it appropriately. For example as in my earlier proposal. -- --Per Bothner per@xxxxxxxxxxx http://per.bothner.com/