Question

If I have a custom method like this:

- (void)myMethod:(id)myArgument 
{
      //do something with myArgument
}

is myArgument guaranteed to stick around for the execution of that method if i don’t take ownership of it?

EDIT
Let me elaborate further. Say i call this somewhere:

[self myMethod:_myIvar];

and then somewhere else, while myMethod is executing, someone calls this:

[_myIvar release];

does that effect the argument within myMethod?

END EDIT

Looking through documentation/sample code, i rarely see a [myArgument retain] or [myArgument copy] at the top of a custom method. So is it unnecessary?

Thanks!

Answer 1

If myArgument is a valid object at the start of myMethod, and you do nothing to cause it to be deallocated or otherwise collected during the method, then yes, myArgument will still be a valid object at the end of the method.

EDIT in response to question’s edit: let’s talk about object ownership in this case. When you’re passed myArgument as an argument to your function, you – in the context of myMethod – don’t declare an interest in the object by retaining or copying it. In such a situation, the only owners of the object exist outside myMethod.

If the last owner of myArgument decides that it’s done with myArgument and releases it, then it’s entirely possible that myArgument goes away – there are no strong references to the object any more that would keep it around. It may be valid, depending on threading concerns, autorelease pools involved, and a whole host of other issues, but that’s a dangerous game to play. If there’s even a possibility that such a situation might happen, you should explicitly declare an interest in myArgument within myMethod, as recommended by the Memory Management Policy.