Re: Keep compiler silence (clang 10, implicit conversion from 'long' to 'double' )

Hello Tom,

Thank you for replying.

On Wed, Nov 6, 2019 at 12:04 AM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>
> Yuya Watari <watari(dot)yuya(at)gmail(dot)com> writes:
> > The added macro FLOAT8_FITS_IN_INT32() does not check NaN explicitly,
> > but it sufficiently handles the case.
>
> Really? I don't think anything is guaranteed about how a NaN will
> compare when using C's non-NaN-aware comparison operators.
>
> My thought about this was to annotate the macros with a reminder
> to also check for NaN if there's any possibility that the value
> is NaN.

I agree with your opinion. Thank you for pointing it out.

If the platform satisfies IEEE-754 standard, all comparisons (except
for "not equals") between NaN and other floating values are "false".
[1] In this case, the proposed FLOAT8_FITS_IN_INT32() macro handles
NaN.

[1] https://en.wikipedia.org/wiki/NaN#Comparison_with_NaN

However, this behavior depends on the platform architecture. As you
have said, C language does not always follow IEEE-754. I think adding
explicit checking of NaN is necessary.

I modified the patch and attached it.

Best regards,
Yuya Watari
NTT Software Innovation Center
watari(dot)yuya(at)gmail(dot)com