CKTIS / C
signbit — checking a floating-point value's sign
The signbit macro from <math.h> checks whether a floating-point value has a negative sign, including zeros, infinities, and NaNs. A nonzero result need not equal 1.
Version: C23; repository compiler flag, production runtime not attested
Syntax
#include <math.h>
int signbit(real-floating x);signbit is a macro from <math.h>, not an ordinary function. In the synopsis, real-floating means a real floating type, such as float, double, or long double; it is not a C type name.
The int result is nonzero if and only if the sign is determined to be negative; otherwise it is zero. Use !!signbit(x) to obtain exactly 0 or 1, rather than comparing the result with 1.
This differs from x < 0.0: a supported negative zero has a negative sign but is not less than zero. The macro also covers infinities and NaNs. An unsigned zero is treated as positive; for other unsigned values, the result is implementation-defined. Here, “unsigned” concerns the floating-point value, not the type unsigned int.
Examples
#include <math.h>
#include <stdio.h>
int main(void)
{
printf("%d %d %d\n",
!!signbit(-2.0),
!!signbit(0.0),
!!signbit(2.0));
return 0;
}Result
1 0 0
The displayed code and declared result were verified in an isolated environment using the method recorded in the verification evidence for this exact version.