Consider not using _ecvt_s

[alabuzhev]

• "The procedure entry point _ecvt_s could not be located in the dynamic link library msvcrt.dll"

This happens on Windows XP if compiled with MinGW, as XP's msvcrt.dll is too old and doesn't include that function, and MinGW doesn't use the static runtime library for copyright reasons.

_ecvt_s is used in implementation of bool isnegative(double x) ("Portable version of signbit").
Probably it's better to just use signbit if it's available. MinGW defines this macro in math.h.

[vitaut]

isnegative already uses signbit by default and I'd expect the call to _ecvt_s to be optimized away if signbit is available. Do you get this error when compiling without optimizations?

[alabuzhev]

http://www.cplusplus.com/reference/cmath/signbit/

In C, this is implemented as a macro that returns an int value. The type of x shall be float, double or long double
In C++, it is implemented with function overloads for each floating-point type, each returning a bool value.

So, if (const_check(sizeof(signbit(x)) == sizeof(int))) is actually if (false).

I compile with -O3, so I presume it will be optimised away indeed, just not the branch you expect.

[vitaut]

Good catch. Does changing

    if (const_check(sizeof(signbit(x)) == sizeof(int)))

to

    if (const_check(sizeof(signbit(x)) == sizeof(int)) || const_check(sizeof(signbit(x)) == sizeof(bool)))

solve the problem?

[alabuzhev]

Yes, with that change I can't find any traces of _ecvt_s in the binary and it works.

[vitaut]

Fixed in def6874.