Return true if the specified floating point number has its sign bit set (i.e., it is negative), and false otherwise. This function is equivalent to the signbit macro in C99. Note that this function will return true for some NaNs, even though the concepts of negative and positive do not apply to NaNs.