为什么浮点数打印为负零？

Question

Problem

Sometimes the floating point number of value zero is being printed out as negative zero even though inequality checks show that the number is non-negative.

Application

I'm trying to log out the transformation matrix of a UIView in a human readable format. The numbers should be arranged like this:

- (void) describe
{
    NSLog(
        @"|%@%.2f %@%.2f 0.00|\n|%@%.2f %@%.2f 0.00|\n|%@%.2f %@%.2f 1.00|",
        self.transform.a < 0 ? @"" : @" ",
        self.transform.a,
        self.transform.b < 0 ? @"" : @" ",
        self.transform.b,
        self.transform.c < 0 ? @"" : @" ",
        self.transform.c,
        self.transform.d < 0 ? @"" : @" ",
        self.transform.d,
        self.transform.tx < 0 ? @"" : @" ",
        self.transform.tx,
        self.transform.ty < 0 ? @"" : @" ",
        self.transform.ty
    );
}

I place spaces in front of numbers that are non-negative to match the width of negative numbers. In some cases, the numbers line up:

| 0.00 -1.00 0.00|
| 1.00  0.00 0.00|
| 0.00  0.00 1.00|

After applying a few rotations to the view, the numbers don't line up anymore:

| -0.00 -1.00 0.00|
| 1.00  -0.00 0.00|
| 0.00  0.00 1.00|

In the above case, an NSLog of self prints out the transform as:

transform = [-0, -1, 1, -0, 0, 0];

The problem is zero values are logged out as negative zero, but being detected as non-negative. If I can understand how this mythical negative zero can even exist, then I can solve my formatting problem.

Answer 1

[Answer moved from comment]

floating point values have mantisa,exponent and sign separate

• so they are capable of storing +0 and -0 which is the same ...
• the sign is leftover from last operation with such value ...