//Parent.m
#import "Parent.h"
#implementation Parent{
dispatch_block_t _block;
NSTimer *_timer;
}
- (instancetype)init
{
self = [super init];
if (self) {
[self commonInitialization];
}
return self;
}
-(void)commonInitialization{
__unsafe_unretained typeof(self) weakSelf=self;
//__weak typeof(self) weakSelf=self; the same conculsion
//apple doc:The object to which to send the message specified by aSelector when the timer fires. The timer maintains a strong reference to this object until it (the timer) is invalidated.
_timer=[NSTimer timerWithTimeInterval:0.5 target:weakSelf selector:#selector(sayHello) userInfo:nil repeats:YES];// make weakSelf retain count +1
_block=^{
__strong Parent *parent=weakSelf;//also make weakSelf retain count +1,but the actual is that this wont make self retain count +1
[parent sayHello];//
};
// my question is why weakSelf can make _block wont retain self,but _timer will retain self,it look like contradictory
}
-(void)sayHello{
NSLog(#"hello");
}
-(void)dealloc{
NSLog(#"Parent instance can dealloc");
}
#end
In _block I retain weakSelf again like NSTimer will retain target:weakSelf.
My question is why __unsafe_unretained or __weak can make _block unretained self but _timer not. It looks contradictory.
Retain cycles with blocks happen when a class instance has a strong reference to the class and the block, in turn, has a strong reference to that instance. By capturing a weak reference to the instance within the block (weakSelf), that captured reference can be set to nil. Making the reference within the block a strong reference only means that while the block is executing, the instance will not be dealloc'd (because the block still has a live reference).
The timer internally maintains a strong reference to the instance, just as the block does while it's executing. The main difference is that the timer is a long-lived object, and unless you invalidate it or nil all your references to it, the timer still exists, along with its strong reference to your instance.
I really do not know, whether I understood you correctly. Do you mean with "makes retained" "retains"?
However, the difference between the two pieces of code is the time of execution and how references are handled:
Keep in mind that not object references are retained, but the object they points to.
A. Timer
{
__weak typeof(self) weakSelf=self;
_timer=[NSTimer timerWithTimeInterval:0.5 target:weakSelf …];
}
With this code you create an additional local var called weakSelf, that does not retain the object it points to. Moreover, the extent ("lifetime") of weakSelf ends with the C-block (not the closure what you called __block), that means with the closing }. So we have:
{
__weak typeof(self) weakSelf=self;
// does not retain self; no problem
_timer=[NSTimer timerWithTimeInterval:0.5 target:weakSelf …];
// weakSelf dies, since it is weak nothing happens.
}
In such a case it is completly meaningless to weakify self:
{
id anotherSelf=self;
// does retain self: +1;
_timer=[NSTimer timerWithTimeInterval:0.5 target:weakSelf …];
// anotherSelf dies: -1
}
This is always balanced, because ARC care about it. No problem.
So why is there a retain cycle? It is "inside the timer." According to the documentation:
The object to which to send the message specified by aSelector when the timer fires. The timer maintains a strong reference to this object until it (the timer) is invalidated.
Therefore, lets go back to your example:
{
__weak typeof(self) weakSelf=self;
// does not retain self;
_timer=[NSTimer timerWithTimeInterval:0.5 target:weakSelf …];
// timer retains the object, weakSelf points to: +1.
// self retains the timer: +1
// result: retain cycle
// weakSelf dies, since it is weak nothing happens.
}
What -timerWithInterval… does, does not depend of the strength of weakSelf. The method does not even see the strength of the reference. It retains an object that the argument points to. Period.
B. Block
Having a block it is different:
{
__weak typeof(self) weakSelf=self;
// does not retain self; no problem
_block=^{
…
};
// weakSelf dies, since it is weak nothing happens.
}
As you can see, there is no problem. Why can be there a retain cycle? This is quite easy: Having a reference inside the block, the referred object is retained when the block is created (similar to timers) and the reference is strong(different to timers):
{
__weak typeof(self) weakSelf=self;
// does not retain self; no problem
_block=^{
… self … // the object self points to is retained, because self is strong. +1
… weakSelf … // the object weakSelf points to is not retained.
…
};
// weakSelf dies, since it is weak nothing happens.
}
This reference lives as long as the block lives. You have to read it correctly: It is like the reference itself has a longer lifetime. Therefore it depends of the strength of the reference.
So there is a big difference, whether you use weakSelf or self (or any other weak or strong reference.)
What is done inside the block …:
_block=^{
__strong id strongSelf=weakSelf;
};
… is meaningless, because this is done, when the block is executed and the local strongSelf will immediately lose its extent. (Again with the closing }.
ARC and block are all compiler work, compiler will add retain release and it will translate block to struct.
NSTimer will hold a strong reference to target, so it will send retain method (it may call objc_retain(id value) in ARC) to weakSelf, self hold timer as a instance variable, retain cycle happens.
Let's take about the block, by default block will retain any NSObject it captures when the block is move from stack to heap.
the _block(in your case) is a instance variable, when it is assigned a block, the block will be send a copy method, so it will move from stack to heap.
But if object it captured is with __weak or __unsafe_unretained attribute, it won't send retain method to it. The rules are defined by compiler, so it works.
If you want to learn the detail, check the source code runtime.c, you may also need to assemble your code because runtime.c doesn't have ARC code.
If you were confused about this line of code __strong Parent *parent=weakSelf , check this answer.
Related
Within blocks, do I need to pass some version of self to helper methods?
In this example, publicMethod() is calling _privateHelper using strongSelf.
But, _privateHelper() is calling _secondPrivateHelper() using self. Is this fine or should I pass down strongSelf to the helpers?
- (void)publicMethod
{
auto __weak weakSelf = self;
dispatch_sync(dispatch_get_main_queue(), ^{
auto __strong strongSelf = weakSelf;
if (!strongSelf) {
return;
}
[strongSelf _privateHelper];
});
}
- (void)_privateHelper
{
[self _secondPrivateHelper]; // <-- what version of `self` is sending the message here? I'm assuming the copied `strongSelf` from above.
}
- (void)_secondPrivateHelper
{
NSLog(#"Hello, World");
}
TLDR
When to use strong or weak depends on how you use the block. In your code strong and weak does not really matter but I will give an example where it matters and explain how to then use weak.
I think you are considering the wrong things here. In your example all pointers to self can (and should) be strong. You do not need to dance the dance here and you should consider using weak only when you create retain cycles which I will discuss in a moment.
In your case note that when you are inside the block you want a strong pointer to self. If you have a weak pointer the block becomes superfluous when self is nil. To get a strong pointer simply change your block to
... ^{
[self _privateHelper];
});
rather than the way you do it. Now (and in your original code) you have (and should have) a strong pointer to self inside both your _privateHelper and _secondPrivateHelper.
But this is not the issue.
Retain cycle
Suppose you have an ivar
#property (strong,nonatomic) void ( ^ block )( void );
Only now do you have potential trouble and you probably need to use weak pointers. The reason being that if somewhere in your code you have
self.block = ^{ [self _privateHelper]; };
the block will retain a strong pointer to self. At the same time, self is keeping, through the ivar, a strong pointer on the block. Can you see the retain cycle? Neither the block nor self will ever be released automatically as both are keeping strong references to the other.
Fixing it
You can change the ivar to weak in which case you break the cycle. If the block goes out of scope the weak ivar goes to nil and there is no retain on self.
You could also do
self.block = ^ { [weakSelf _privateHelper]; };
inside the block to break the cycle. Now if self goes out of scope the block no longer has a hold on it and again there is no retain cycle on self.
Both these fixes are used - a lot of ivars will be weak in a UI environment to prevent the e.g. label keeping a strong reference on the vc while the vc also keeps a strong reference on the label.
When working with blocks you typically use weak pointers to self inside the block only if you potentially have a retain cycle. This is often not an issue and in your code also is not an issue as self does not retain the block. So in general your helper blocks could probably get away with strong pointers to self. You only need to dance the dance if somehow self and the block are creating a retain cycle.
Finally, a third way to fix it is to set the ivar to nil yourself. So you could have code that looks like this
if ( self.block )
{
self.block (); // Execute block
self.block = nil;
}
This also makes sense sometimes e.g. when you want to retain the block and self until the block executes. Here you release the ivar manually by setting it to nil after which it no longer retains self and the cycle is broken albeit manually.
You could even do this inside the block, e.g. change the block to
self.block = ^ { [self _privateHelper]; self.block = nil; /* release when done */ };
When I pass self as an argument to a function, does it make any difference if I weakify it first?
For example:
__weak __typeof(self) weakSelf = self;
[self.someObject doWorkWithDelegate: weakSelf];
In the body of the doWork function it assigns it to a strong property
#property (strong) Foo* fooDelegate;
(I know this is terrible, don't get sidetracked).
This will still be a strong reference and cause a memory cycle despite theh fact that I "weakified" the reference first, correct?
Bonus Question: How can I check this myself?
It is the variable weakSelf which is weak. That is, ARC does not emit a retain when a value is assigned to it, nor a release when it goes out of scope (or is assigned a different value).
Since the fooDelegate property is strong, assigning to it releases any old value it may have had and retains the newly-assigned value.
So, yes, that will be a strong reference. It's not clear from what you posted whether it will constitute a retain cycle. The name of the method suggests that the delegate will be cleared (thus released) after the work has been completed. In that case, it's not a problem. It would only be a problem if self.someObject maintained the strong reference until it itself was released and self maintained a strong reference to someObject until self was released.
An important aspect of the original question could be further clarified. "When I pass self as an argument to a function, does it make any difference if I weakify it first?"
Note that in the example code block:
__weak __typeof(self) weakSelf = self;
[self.someObject doWorkWithDelegate: weakSelf];
ARC will perform a retain on each of the objects passed, including the receiver (i.e. self.someObject and weakSelf), and then a release on each when complete, allowing weak objects to be safely used for the lifetime of the called method. So the answer would be no, it doesn't make a difference to make it weak first because it will be strong for the duration of the call.
If the weak variable was simply referenced within a code block callback from the method, then it will still be weak within that scope.
So this is a common pattern to use a weakSelf variable within a block callback to call another method, foo, which then can safely use self because it is no longer weak during that execution of foo. But if self goes away before the callback is executed, then [weakSelf foo] will simply never be called because weakSelf has gone nil.
__weak __typeof(self) weakSelf = self;
[self.someObject doWorkWithCallback:^{
[weakSelf foo];
}];
- (void)foo
{
// Safely use self here.
[self doMoreWork];
}
I want to use an ivar in a method that is being called from a block. Is it fine to call the ivar directly from it without causing a retain cycle?
I have seen other questions related to calling properties- saying that we 'can' use self. Just wanted to make sure about ivars, can we call ivar directly or do we need to use self->_myInstanceVariable explicitly?
__weak MyClassName *weakSelf = self;
[MyBlockCreator myBlock:^{
MyClassName *strongSelf = weakSelf;
[strongSelf doSomething];
}];
- (void) doSomething {
_myInstanceVariable = someObject
// is it fine to use the ivar directly above?
// or do we want use self explicitly?
// like self->_myInstanceVariable
}
In this case you don't need to make the reference weak/strong at all, the only situation in which you can cause a retain cycle is if self references a block and a strong version of self is captured within that block. Since self doesn't retain the block in question, there would be no retain cycle even using a strong reference.
[Foo doWithBlock:^{
bar.baz = #"Hello World"; // Totally fine without weak/strong since bar doesn't retain the block here.
}];
In situations where self holds a strong reference to a block and the block refers to self, you need to make the reference weak outside the block then make it strong again inside.
__weak id weakSelf = self;
[self doWithBlock:^{
id strongSelf = weakSelf;
strongSelf.baz = #"Ok";
}];
If the code inside a block calls a method, will a retain cycle exist if that method references self? In other words, does all code downstream of a block need to use the weakSelf/strongSelf pattern?
For example:
__weak __typeof__(self) weakSelf = self;
Myblock block = ^{
[weakSelf doSomething];
};
. . .
- (void)doSomething
{
self.myProperty = 5; // Is this ok or does it need to use a weakSelf?
}
Objective-C is not scoped like you suggest, namely, you don't have access to weakSelf from within -doSomething. Furthermore, as you are calling -doSomething on weakSelf, "self" within that call is actually referring to the same object that weakSelf is.
In short, no, you shouldn't, you can't and you shouldn't.
Retain cycle will be triggered only if you retain self inside the block. Otherwise it will just throw a warning only.
This is fine you can use this. Because block retains every vars used inside, so retain cycle would be like
Self would retain block
If block retains self then
Self would again retain block
block would retain self, so cycle goes on
The thing you are doing in method is just message passing. Everytime block is called a message would be sent to self to doSomething. And you can retain self in doSomething method it wont trigger retain cycle because this method dont have cycle loop to self. Hope you understand :)
- (void)doSomething
{
self.myProperty = 5; // Is this ok or does it need to use a weakSelf?
}
you can do this to get rid of retain cycle problem.
[self class] __weak *weakSelf = self;
self.completionBlock = ^{
[self class] __strong *strongSelf = weakSelf
[weakSelf doSomething];
};
Here is the block in question. Im wondering if declaring strongSelf outside of the block will introduce a retain cycle in anyway. The reason for setting up the strongSelf outside the block is that I'm worried that weakSelf will be deallocated before the completion block.
__weak CollectionView *weakSelf = self;
CollectionView *strongSelf = weakSelf;
[self
performBatchUpdates:
^{
strongSelf->_cellSet = [strongSelf performDeletions:[strongSelf deletionsTowards:destinationSet from:strongSelf->_cellSet] from:strongSelf->_cellSet];
[(CollectionViewLayout *)strongSelf.collectionViewLayout setCellSet:strongSelf->_cellSet];
}
completion:
^(BOOL finished)
{
[strongSelf insertTowards:destinationSet];
}];
What you do is just pointless. weakSelf is a weak reference to self. strongSelf is a strong reference to self. It is captured by the block, so the block now also has a strong reference to self (that's the difference when you use weakSelf in the block: When the block captures weakSelf, the block has a weak reference to self). Instead of using strongSelf in the block, it would be much easier to just use self - it will be exactly the same thing. You might confuse the compiler enough to not give you any warnings when you use strongSelf where you would have got a warning for using self (if that warning was appropriate).
Whether you cause a retain cycle: Depends on the code. If a copy of the two blocks is stored in the object self and not set back to nil when the completion blocks are called, then you have a retain cycle. If the two blocks are not stored anywhere then the strong reference to self goes away eventually.
The method that actually works well is to assign weakSelf to a strong reference strongSelf inside each block. And then the first thing that you do is to test once whether strongSelf is nil. In that case you know that self has been deallocated, and you need to think about what actions you would want to do in that case.