In this question, I asked about the following code and retain cycles:
__weak Cell *weakSelf = self;
NSBlockOperation *op = [NSBlockOperation blockOperationWi
I can't tell you the reason for the retain-cycle in your second example, because I don't know MBItem
, but there are two different usage patterns with blocks.
If you expect your block to execute in any case, then you can just use self
in the block:
[startSomeOperationWithCompletionBlock:^{
[self doSomeThing];
}];
The block retains a reference to self
, so that self
is not deallocated before the block is executed. But after the block has executed, this reference (and the retain cycle) is gone.
If you possibly want that self
is deallocated before the block has executed,
or if it is possible that the block will not be called at all,
then you have to use a weak reference and check the value inside the block:
__weak MyClass *weakSelf = self;
[startSomeOperationWithCompletionBlock:^{
MyClass *strongSelf = weakSelf;
if (strongSelf) {
[strongSelf doSomeThing];
}
}];
The block does not retain self
in this case, so that self
can be deallocated. In that case, weakSelf
is set to nil
automatically. Therefore, if the block is executed finally, you have to check first if weakSelf
is still valid. (Or you can just use it, because sending messages to nil
is a no-op.)
Assigning a strong reference strongSelf
inside the block prevents self
from being deallocated while the block is executing.