Question

When using ARC for iOS, is there any difference between the following?

@property (strong, nonatomic) NSObject *someProperty;
...
@synthesize someProperty;

//and then in the init method, either:
self.someProperty = aProperty;

//or
someProperty = aProperty;

I know that without ARC, self.someProperty is actually calling the synthesized setter method which sends a retain message to the object. But now with ARC, does it matter if I use dot notation for setting a property like this?

More generally, does ARC truly mean that I don’t have to worry about reference counts at all? Or are there certain situations in which the way I wrote my code could cause ARC to make a mistake?

Answer 1

The difference is the same as in the case without ARC: by using dot notation, you are calling the synthesized setter, and by assigning directly to the ivar, you are going around the setter method.

Under ARC, there are no differences in memory management between the two options but you should still make a conscious decision between the two options: assigning directly to the ivar bypasses KVO, for example, while going through the setter method is slightly slower but probably safer in most cases, e.g. when you later decide to make the property atomic or override the setter.

Personally, I would always use the property notation self.abc = ...; except possibly in init where it is often desirable to bypass KVO. In short, use the same reasoning you used before ARC.