开发者

iPhone alloc and release

i am new in ObjC and iPhone.

I downloaded an example of data sharing between multiple views. The basic approach is to create an data model object in the base UIApplication and get/set data from/to it. So in the init method i saw the following code:


- (id) init;
{
    self.theAppDataObject = [[ExampleAppDataObject alloc] init];
    [theAppDataObject release];
    return [super init];
}
开发者_JAVA百科

And after that, using delegate we can access this object.


id theDelegate = (id) [UIApplication sharedApplication].delegate;
    ExampleAppDataObject* theDataObject;
    theDataObject = (ExampleAppDataObject*) theDelegate.theAppDataObject;

So, my question is in the first code example. Why do we need to alloc memory for the theAppDataObject object, and immediately after that - release the object? Why don't we get nil when accessing this object later?

10x


I assume theAppDataObject is declared as @property (retain). Therefore, when setting the object by doing self.theAppDataObject (or [self setTheAppDataObject:]), this property will retain the ExampleAppDataObject. Therefore, you can release it afterwards.

When you alloc and init the ExampleAppDataObject, it's retain count goes up to 1. When you set the AppDataObject to this ExampleAppDataObject, it sends retain to it, so the retain count goes up to 2. You can then release your own ownership of the object; it won't get deallocated because theAppDataObject still has ownership.

If that makes sense.


That depends on how theAppDataObject property is defined. If it provides a retaining setter-accessor the retain count of the appDataObject will flip to 2, one more than needed here.

So release one of it.

Better and more understandable way would be to write

if ( (self = [super init]) ) {
    ExampleAppDataObject *myAppDataObject = [[ExampleAppDataObject alloc] init];
    self.theAppDataObject = myAppDataObject;
    [myAppDataObject release];
}

return self;


Iphone uses reference count based memory management model..See this tutorial first and then apple's technical documentation... theAppDataObject is a property (see the use of self. theAppDataObject) which should be retained for the above code to work..Any object which is retained should have a bonus plus 1 retain count...An object only gets released when its retain count gets to zero..


first things first: that code sample is terrible.

- (id) init
{
    // assign self. super may return another address
    self = [super init];
    // then check for nil
    if (self != nil) {
            // then assign the ivar directly because you should
            // not invoke methods of partially constructed objects
        theAppDataObject = [[ExampleAppDataObject alloc] init];
    }
    // then return the address super returned
    return self;
}

now for your questions:

Why do we need to alloc memory for the theAppDataObject object, and immediately after that - release the object?

self.theAppDataObject calls through the setter, which retains, copies, or assigns theAppDataObject. in this case, we could assume it is highly likely to be retained.

Why don't we get nil when accessing this object later?

release does not set a pointer to nil. it sends a message to the object which then decrements the retain count (in the typical case). what you might have expected in this case is an object which has been deallocated. that does not happen when the argument is retained because the reference count does not reach zero in the program you have shown. the object is still alive because it's been retained and the address of the object stored when the setter (self.theAppDataObject = arg) is called.

0

上一篇:

下一篇:

精彩评论

暂无评论...
验证码 换一张
取 消

最新问答

问答排行榜