Ranier Vilela <ranier(dot)vf(at)gmail(dot)com> writes:
> Yes, but the comment it does not clarify that the return of the variable
> "const_one" is intentional, instead of "const_zero".
I'm not sure which part of "NaN ^ 0 = 1" doesn't clarify for you that
the intended result is 1.
Even without the comment, if you'd bothered to run the regression tests
you'd have noted a failure of a test case clearly intended to test
exactly this behavior.
regards, tom lane