[Cfp-interest 2601] Re: draft CFP response for NB comments and N3071

[Jim Thomas]

> On Jan 7, 2023, at 11:34 AM, Fred J. Tydeman <tydeman at tybor.com> wrote:
> 
> On Fri, 6 Jan 2023 12:10:25 -0800 Jim Thomas wrote:
>> 
>> Please review the paper and send any corrections or additions ASAP.
> 
> Do the examples that involve "0.1" also depend upon FLT_EVAL_METHOD?
> If long double has more precision than double,
> constexpr double d = 0.1;
> would have a change in value if FLT_EVAL_METHOD were 2.
> But,
> constexpr double d = (double)0.1;
> would always be safe.

The only example (in our proposed change) with “0.1” is

constexpr float _Complex fc2 = 0.1; 	// constraint violation, unless double has
								// the same precision as float

If no objections, I’ll append to the comment “and is evaluated with no extra precision”

> 
> It might help to add to your examples:
> constexpr double _Complex dc1 = DBL_SNAN; // ok???
> constexpr double _Complex dc2 = CMPLX(DBL_SNAN, 0.);  // ok
> constexpr double _Complex dc3 = CMPLX(0., DBL_SNAN); // ok

If no objections, I’ll add

constexpr double _Complex dc1 = DBL_SNAN;		// constraint violation
constexpr double _Complex dc2 = CMPLX(DBL_SNAN, 0.);	// ok
constexpr double _Complex dc3 = CMPLX(0., DBL_SNAN);	// ok

- Jim

> 
> 
> 
> 
> ---
> Fred J. Tydeman        Tydeman Consulting
> tydeman at tybor.com      Testing, numerics, programming
> +1 (702) 608-6093      Vice-chair of INCITS/C (ANSI "C")
> Sample C17+FPCE tests: http://www.tybor.com
> Savers sleep well, investors eat well, spenders work forever.