I have seen sample source code around that uses different ways of releasing/dealloc'ing objects, so I was wondering which way is deemed the "best" one. A few options on -dealloc:
1) Release objects
- (void)dealloc {
[obj1 release];
[obj2 release];
[super dealloc];
}
2) Set objects to nil
- (void)dealloc {
self.obj1 = nil;
self.obj2 = nil;
[super dealloc];
}
3) De-allocate objects directly
- (void)dealloc {
[obj1 dealloc];
[obj2 dealloc];
[super dealloc];
}
Which way is the best one? Pros and cons for each?
Method 1 is the only recommended method. It's also good practice to set them to nil AFTER you've released them.
Method 2 only works for properties that manage their own object/value retaining, so it's not universally applicable. And if you implement your own setter method that performs other actions when the property changes, you may get undesired side effects by calling it in [dealloc].
Method 3 violates the reference-counting principle because it will deallocate the objects even if something else has a [retain] hold on them, so when the other objects access them next, your program will crash. You're never supposed to call [dealloc] directly -- let the runtime call it when the last owner calls [release].
If obj1 and obj2 are properties using #synthesize-d accessors, then method 1 and method 2 are equivalent. If either of those two conditions is not true then the effect is rather different. If they're not properties, then method 2 just sets the pointers to nil without doing anything that would deallocate the objects. And if they are properties but you've implemented your own setter methods, method 2 calls those methods with "nil" as the argument. Whether that's equivalent to method 1 or not depends on how you've implemented those methods. If your custom setters do anything significant when the argument is nil, method 2 would ensure that this code executed.
As Marco said, method 3 is just plain wrong. If you've seen sample code doing that then the author may not actually know what they're doing.
Related
This question already has answers here:
Closed 10 years ago.
Possible Duplicate:
Dealloc method in iOS and setting objects to nil
As to property dealloc in object c, i have seen different kinds of forms. Which of following is best/correct
//Kind1:
- (void)dealloc
{
[_property release];
[super dealloc];
}
//Kind2:
- (void)dealloc
{
self.property = nil;
[super dealloc];
}
//Kind3:
- (void)dealloc
{
[_property release]; _property = nil;
[super dealloc];
}
Your second option is inadvisable. Property setters may have side-effects and generally aren't implemented with the consideration that the object may be half torn down already. In general you should avoid any method calls to self (or super) from dealloc, other than [super dealloc]. You should also avoid non-trivial calls (i.e. anything but release) on other objects, as you may have circular references that can lead back to references to your half-deallocated object.
The first option is perfectly sufficient. Note that if you use ARC it is redundant. Using ARC is your best bet, invariably, as it's not only safer but faster.
The third option is controversial. Some people argue that it makes your program more resilient to errors (since references after dealloc may end at the zeroed instance variable, rather than bouncing through it and seg faulting, or worse). They also argue that it ensures that if you do run a method which tries to use the value, it'll probably fail gracefully (again, rather than dereferencing what is a dangling pointer at that point). But if you subscribe to my first point on avoiding this to begin with, it becomes somewhat moot. And my experience is that relying on that behaviour is a bad idea - even if it's a clean design to begin with, it's easy to forget about it and end up changing the code some time later, in a way that breaks it.
And those that dislike the third option also argue that it hides errors. Some go as far as explicitly overwriting not with nil but rather an obviously bogus value (e.g. 0x0badf00d) to make it clearer if and when a dangling pointer is dereferenced.
Kind1 is more than enough;
It is usually not a good idea to use 'self' on alloc and dealloc. coz self will call setter/getter method, and if u have custom setter/getter method its can cause trouble.
You should never call self. in dealloc.
Once you have release-ed then there is no advantage whatsoever of making it nil.
So, first one is the correct way to use dealloc.
OK, looking at this:
Apple docs: Declared Properties
If you scroll down to dealloc it reads:
"Typically in a dealloc method you should release object instance variables directly (rather than invoking a set accessor and passing nil as the parameter), as illustrated in this example:"
- (void)dealloc {
[property release];
[super dealloc];
}
"If you are using the modern runtime and synthesizing the instance variable, however, you cannot access the instance variable directly, so you must invoke the accessor method:"
- (void)dealloc {
[self setProperty:nil];
[super dealloc];
}
Now, I must own at least 15 to 20 books on iOS development. I can't say that I have ever seen any code in these books proposing that one do anything other than:
[someproperty release];
Is there a compelling reason to edit a bunch of files of code that works perfectly well to adopt Apple's recommendation? How about future work? Or, are they pretty much equivalent?
Apple's document appears to apply only if you have synthesize-by-default turned on. If you have it turned off, which seems to be the default case, you need #synthesize, which gives you access to variables directly.
Personally, I do:
#synthesize someProperty=_someProperty;
just so I can then do:
- (void)dealloc {
[_someProperty release];
[super dealloc];
}
See Jeff LaMarche's blog for more.
On the modern runtime, you no longer have to declare instance variables for properties, nor do you have to write #synthesize in the implementation. If your code already has the instance variables declared, then there's no reason to replace code that calls release with a call to a set accessor (the first quote you pasted even states that).
What is the recommended way of doing this. Should I call super dealloc first or last or doesn't it matter?
- (void)dealloc
{
[super dealloc];
[orderNumber release];
[orderDate release];
}
Also when it comes to overriding methods like didViewLoad - should I call super first or last?
Always call [super dealloc] last or you might easily come into trouble because you're working on a stale object.
With didViewLoad you normally call it before your own code as you want the standard initialization stuff executed before. I've seen examples in Apple's code that don't call the super implementation at all, though, so maybe there's not much going on anyway.
In this case call the super after you have released all your properties/iVars. For viewDidLoad/willAppear/etc. I usually call the super first. The order matters when your custom class is relying on an object that is created by the super. For the default viewDidLoad this is not the case so it is preference(I believe).
There is no general rule - you chose to override the method, what does it do? Do you want it to happen before or after your custom implementation?
didViewLoad doesn't appear to be a real method.
We know that [super dealloc] destroys the current object completely, so any code that comes after it is wrong. So, in this case, call [super dealloc] last.
The pointers orderNumber and orderDate are held inside your object.
[super dealloc] deallocates your object (aka self).
Once you deallocate your object you must not rely on the things inside it (e.g. orderNumber) having the values they did before you deallocated it.
Therefore, deallocate the members before deallocating the self object.
The opposite holds true for init functions - you can't initialise the pointers until after your object is constructed, so [super init] comes before you initialise the members.
Regarding viewDidLoad (et al), you do whatever works. If you have stuff you want to happen before the superclass does its thing, then you do it before you call the superclass method, and likewise for stuff you want to happen afterwards.
If you don't know whether your code should run before or after, then it probably doesn't matter.
Wondering if someone with experience could possibly explain this a bit more. I have seen examples of...
[view release];
view = nil;
....inside the (void) dealloc.
What is the difference and is one better then the other?
What is the best way?
When doing retainCount testing I have personally seen nil drop a count from 3 to 0 for me, but release only drops it from 3 to 2.
What you have seen is probably these:
1) [foo release];
2) self.bar = nil;
3) baz = nil;
Is releasing the object, accessing it through the instance variable foo. The instance variable will become a dangling pointer. This is the preferred method in dealloc.
Is assigning nil to a property bar on self, that will in practice release whatever the property is currently retaining. Do this if you have a custom setter for the property, that is supposed to cleanup more than just the instance variable backing the property.
Will overwrite the pointer baz referencing the object with nil, but not release the object. The result is a memory leak. Never do this.
If you are not using properties (where self.property = nil will also release an object) then you should ALWAYS follow a release by code that sets the reference to nil, as you outlined:
[view release]; view = nil;
The reason is that it avoids he possibility that a reference can be used that is invalid. It's rare and hard to have happen, but it can occur.
This is even more important in viewDidUnload, if you are freeing IBOutlets - that's a more realistic scenario where a reference might go bad because of memory warnings unloading a view, and then some other code in the view trying to make use of a reference before the view is reloaded.
Basically it's just good practice and it will save you a crash at some point if you make it a habit to do this.
#bbullis22 you have seen the restain count drop from 3 to 0 because you set the reference to nil. then you asked for the retaincount of 'nil' which is zero. however, the object that used to be referenced has the same retain count - 1 (due to setting the reference to nil).
using release, the reference still references the same object, so that's why you see the retain count drop from 3 to 2 in this situation.
As far as usage inside your code, in your dealloc you don't need the assignment to the property, releasing is all you need to do.
- (void)dealloc {
[myProperty release]; // don't need to assign since you won't have the object soon anyway
[super dealloc];
}
I think using both is kind of safety net. With only release in place you could run in problem if you screwed reference counting management. You would release an object, giving its memory back to system but pointer would be still valid.
With nil you are guaranteed that program will not crash since sending message to nil does nothing.
Example:
- (NSString*) title {
return [[title retain] autorelease];
}
The setter actually retained it already, right? and actually nobody should bypass the Setter... so I wonder why the getter not just returns the object? It's actually retained already. Or would this just be needed in case that in the mean time another objects gets passed to the setter?
From here http://www.macosxguru.net/article.php?story=20030713184140267
- (id)getMyInstance
{
return myInstanceVar ;
}
or
- (id)getMyInstance
{
return [[myInstanceVar retain] autorelease] ;
}
What's the difference ?
The second one allows the caller to get an instance variable of a container object, dispose of the container and continue to play with the instance variable until the next release of the current autoreleased pool, without being hurt by the release of the instance variable indirectly generated by the release of its container:
aLocalVar = [aContainer getAnInstanceVar] ;
[aContainer release];
doSomething(aLocalVar);
If the "get" is implemented in the first form, you should write:
aLocalVar = [[aContainer getAnInstanceVar] retain];
[aContainer release];
doSomething(aLocalVar);
[aLovalVar release];
The first form is a little bit more efficent in term of code execution speed.
However, if you are writing frameworks to be used by others, maybe the second version should be recommanded: it makes life a little bit easier to people using your framework: they don't have to think too much about what they are doing…;)
If you choose the first style version, state it clearly in your documentation… Whatever way you will be choosing, remember that changing from version 1 to version 2 is save for client code, when going back from version 2 to version 1 will break existing client code…
It's not just for cases where someone releases the container, since in that case it's more obvious that they should retain the object themselves. Consider this code:
NSString* newValue = #"new";
NSString* oldValue = [foo someStringValue];
[foo setSomeStringValue:newValue];
// Go on to do something with oldValue
This looks reasonable, but if neither the setter nor the getter uses autorelease the "Go on to do something" part will likely crash, because oldValue has now been deallocated (assuming nobody else had retained it). You usually want to use Technique 1 or Technique 2 from Apple's accessor method examples so code like the above will work as most people will expect.
Compare this code
return [[title retain] release]; // releases immediately
with this
return [[title retain] autorelease]; // releases at end of current run loop (or if autorelease pool is drained earlier)
The second one guarantees that a client will have a non-dealloced object to work with.
This can be useful in a situation like this (client code):
NSString *thing = [obj title];
[obj setTitle:nil]; // here you could hit retainCount 0!
NSLog(#"Length %d", [thing length]); // here thing might be dealloced already!
The retain (and use of autorelease instead of release) in your title method prevents this code from blowing up. The autoreleased object will not have its release method called until AFTER the current call stack is done executing (end of the current run loop). This gives all client code in the call stack a chance to use this object without worrying about it getting dealloc'ed.
The Important Thing To Remember: This ain't Java, Ruby or PHP. Just because you have a reference to an object in yer [sic] variable does NOT ensure that you won't get it dealloc'ed from under you. You have to retain it, but then you'd have to remember to release it. Autorelease lets you avoid this. You should always use autorelease unless you're dealing with properties or loops with many iterations (and probably not even then unless a problem occurs).
I haven't seen this pattern before, but it seems fairly pointless to me. I guess the intent is to keep the returned value safe if the client code calls "release" on the parent object. It doesn't really hurt anything, but I doubt that situation comes up all that often in well-designed libraries.
Ah, ok. from the documentation smorgan linked to, it seems this is now one of the methods that Apple is currently recommending that people use. I think I still prefer the old-school version:
- (NSString *) value
{
return myValue;
}
- (void) setValue: (NSString *) newValue
{
if (newValue != myValue)
{
[myValue autorelease]; // actually, I nearly always use 'release' here
myValue = [newValue retain];
}
}