Float.floatToRawIntBits(b) == negativeZeroFloatBits

I am writing tests for some methods in the Math class of OpenJDK to get coverage (node coverage, edge coverage, etc).

I created a Control Flow graph and found the test paths to get the different types of coverage. One of the paths involves a conditional that has the following statement:

if(( .... && ... && Float.floatToRawIntBits(b)==negativeZeroFloatBits )) I looked up the documentation as well as another stackoverflow Q but am not fully understanding these two parts of the statement.

floatToRawIntBits is defined as: Returns a representation of the specified floating-point value according to the IEEE 754 floating-point "single format" bit layout, preserving Not-a-Number (NaN) values. Not sure what these even means to be honest? Is it saying the number will STAY a Float, even if its NaN (e.g. 0.0/0.0) ?

What about negativeZeroFloatBits? Is this just equivalent to: Float.floatToRawIntBits(-0.0f)

When would the two be equal? When would they not?

(Documentation: http://ift.tt/1SefYme )

(The Math class can be found here: http://ift.tt/2453d1e)

MIN Method:

public static float min(float a, float b) {

    if (a != a)

        return a;   // a is NaN

    if ((a == 0.0f) &&

        (b == 0.0f) &&

        (Float.floatToRawIntBits(b) == negativeZeroFloatBits)) {

        // Raw conversion ok since NaN can't map to -0.0.

        return b;

    }

    return (a <= b) ? a : b;

}