[r6rs-discuss] response to formal comment 8 [equality predicates and NaN]

[Alan Watson]

Per Bothner wrote:
>  > For example, a system with delayed boxing of
>  > inexact real numbers may box the two arguments to eqv?
>  > separately, the boxing process may involve a change of
>  > precision, and the two separate changes of precision may
>  > result in two different payloads.
> 
> I really can't visualize how this could be.  If two unboxed
> flonums have the same precision, representation, and bit-pattern,
> how could any sane implementation box them into two values
> with different precision?

Indeed. I understand that the IEEE spec allows great freedom in this 
area, but does anyone know of real hardware or a real Scheme 
implementation that behaves in this way?

If such a system did exist, an alternative might be to make all NaNs 
eqv?, even if they have different payloads. This would guarantee that 
(let ((x ...)) (eqv? x x)) was true for all numbers.

Regards,

Alan