How can I get the sign-bit of a double?

Question

What is the C# equivalent for C++

bool std::signbit( double arg );

The C++ reference description of this function is as follows: Determines if the given floating point number arg is negative. This function detects the sign bit of zeroes, infinities, and NaNs. Along with std::copysign, this macro is one of the only two portable ways to examine the sign of a NaN.

signbit(+0.0) = false
signbit(-0.0) = true

Answer 1

I'm not sure whether there is a builtin, but this should work:

private static readonly long SignMask = BitConverter.DoubleToInt64Bits(-0.0) ^
    BitConverter.DoubleToInt64Bits(+0.0);

public static bool signbit(double arg)
{
    return (BitConverter.DoubleToInt64Bits(arg) & SignMask) == SignMask;
}

Based on this answer by Jon Skeet.

Answer 2

Reverse-engineered from the machine code in the Microsoft CRT, keeping it fast:

    public unsafe static bool stdsign(double d) {
        ushort* pd = (ushort*)&d;
        return (pd[3] & 0x8000) != 0;
    }

Passes all the tests that I know how to throw at it. Little-endian architecture assumed.