Objective-C Float / Double precision

问题

I was messing around with storing floats and doubles using NSUserDefaults for use in an iPhone application, and I came across some inconsistencies in how the precision works with them, and how I understood it works.

This works exactly as I figured:

{
    NSString *key = @"OneLastKey";
    [PPrefs setFloat:235.1f forKey:key];
    GHAssertFalse([PPrefs getFloatForKey:key] == 235.1, @"");
    [PPrefs removeObjectForKey:key];
}

However, this one doesn't:

{
    NSString *key = @"SomeDoubleKey";
    [PPrefs setDouble:234.32 forKey:key];
    GHAssertEquals([PPrefs getDoubleForKey:key], 234.32, @"");
    [PPrefs removeObjectForKey:key];
}

This is the output GHUnit gives me:

'234.320007324' should be equal to '234.32'. 

But, if I first cast the double to a float, and then to a double it works without fail:

{
    NSString *key = @"SomeDoubleKey";
    [PPrefs setDouble:234.32 forKey:key];
    GHAssertEquals([PPrefs getDoubleForKey:key], (double)(float)234.32, @"");
    [PPrefs removeObjectForKey:key];
}

I was under the assumption that numbers entered without an 'f' at the end were already considered doubles. Is this incorrect? If so, why does casting to a float and then double work correctly?

回答1:

Solved! Turns out my framework method +(void)setDouble:(double)value forKey:(NSString*)key was actually defined as +(void)setDouble:(float)value forKey:(NSString*)key. The value passed was a double but was converted to a float for use in the method. A simple copy and paste issue. Too bad the Objective-C compiler didn't at least throw up a warning like it seems to do for everything else...

来源：https://stackoverflow.com/questions/5408677/objective-c-float-double-precision