I have a NSMutableArray with a few view elements.
I remove the view from the view hierarchy and then check the retain count of the array, it returns 1. I then send [array release] but the code dumps with EXC_BAD_ACCESS.
I see that there are elements in the array but still the code dumps.
Here is what I found during debugging. The array has all the 100 objects present (count on the array returns 100) and the contents when seen on a debugger returns "Out of Scope" for the elements. Also, since the array is in the view hierarchy, the following code reduces the retain count by two:
for (Liv *view1 in viewArray){
NSLog(#"view count = %d", [view1 retainCount]);
[view1 removeFromSuperview];
NSLog(#"view count = %d", [view1 retainCount]);
}
Do not use retainCount
It is useless for this kind of debugging; you are working with views in the framework's view hierarchy. There could be any number of reasons why the retain count goes up or down by 2, 10, or 42.
From the roundabout evidence posted so far, this appears to be a very straightforward memory management issue.
First, use "build and analyze" to have the llvm static analyzer check your code. Fix any problems it identifies.
Next, how is the array allocated? Where do you store it? Post all of the lines of code that declare or manipulate the array.
Finally, as Paul said, turn on zombies and see what happens.
There is an off chance that this isn't a retain/release issue, but there isn't any evidence to indicate that yet.
Did you retain or alloc the array? If not, you shouldn't be sending it a release.
EXC_BAD_ACCESS means you're sending a message to a object that's already been released. You (or some code somewhere) is releasing it prior to the part of your code where you're sending release (or removeAllObjects).
Basically, if your retains and releases are balanced, you won't get this error. We'll need to see some code before we can offer anything more than generic advice.
Using NSLog() and retainCount isn't the easiest way to debug errors of this nature. Assuming you're on Xcode 3.2.6, then try running your code via Run > Run with Performance Tool > Zombies. That should give you a good pointer as to which part of your code is faulty.
Related
I have a problem regarding the memory management when adding objects to a NSMutableArray. Weird thing is that it's all working fine for the first 8 objects I add, but when adding a 9th, the application crashes when retrieving this object.
UploadStatus *status = [[UploadStatus alloc] initWithStatus:[NSString stringWithFormat:#"%#: %d", NSLocalizedString(#"uploadPictureSucceeded", #""), pic_id]
andImageInProgress:nil
imageForSuccess:nil
imageForFailed:nil];
[self.delegate notify:status];
[status release];
This is being done on several places with different texts. But this object contains my status that I display in a UITableView.
In the notify method of the delegate I add the UploadStatus object to the NSMutableArray and I reload the UITableView that shows the objects inside that array.
The first 8 times I add a UploadStatus object to the array and reload the table, it shows correctly. But the 9th time I get the error [CFString retain]: message sent to deallocated instance 0x5c655c0. This error occurs when reloading the table in the cellForRowAtIndexPath method.
Weird thing is that it always shows that the objects inside the NSMutableArray are out of scope like in this screenshot:
Nevertheless if I fetch the item, convert it into the UploadStatus class and get the status from it, it all goes smoothly (for the first 8 objects).
Does anybody have a clue why it goes wrong after adding the 9th UploadStatus object to the NSMutableArray?
Thanks a lot for your help!
The problem is with this code:
[NSString stringWithFormat:#"%#: %d", NSLocalizedString(#"uploadPictureSucceeded", #""), pic_id]
You aren't retaining the string, so it goes away on the next execution of the run loop. You're getting lucky with the first 8. They happen to not get overwritten for some reason, or possibly some other object is retaining them. But the 9th one isn't and you finally see the results of the mistake.
You need for the UploadStatus object to retain that string (and later release it).
I note that you're directly accessing your ivars in this block of code rather than using accessors. This is almost certainly the root of your problem (it is the #1 cause of memory management problems in ObjC). Switch to accessors and most of your memory management problems will go away.
You should also go ahead and run the static analyzer (Build>Analyze). It might shed light. The problem is likely not in the above code; it's somewhere that you're storing something, most likely in an an ivar.
After doing a long profile test, I found that in one of my ".m" file memory leak occurs in the viewdidload section. i checked and the xcode highlighted the part where i have initialized picker arrays with values. my program uses pickers for user input. and i have 3 5 different views in my program. the first is a disclaimer ,the second is a menu where the user can choose the type of calculation he/she wants to do. each calculation requires certain inputs which the user enters from a picker. for eg. one of the view has 5 inputs which are handled by 5 different uipickers with seperate arrays for holding the values. these arrays are initialized with the values in the viewdidload method of that view. here is what i found after running the test:
...................................................................................................
This is my first time developing an app and i'm kinda confused about what to do. Any help would be appreciated.
Objects in objective c have a retain count. If this retain count is greater that 0 when the object goes out of scope (when you stop using it), it leaks.
The following things increase the retain count
[[alloc] init]
new
copy
[retain]
adding an object to an array
adding an object as a child (e.g. views)
There are likely more, but you don't appear to use any others in your code
The following decrease the retain count
[release]
removing an object from an array
if you dealloc an array, all of its objects are released
You should go through your code and ensure each of the retains or additions to an array are matched with a corresponding release. (You can release member variables in the dealloc method).
EDIT: Jeremy made a valid point that my answer doesn't
Once you add an object to an array, it takes ownership and will release the object when it is done with it. All you need to do is make sure you release anything you own according to the memory management rules
There are also autorelease objects, have a look at this example;
-(init){
...
stagePickerArray = [[NSMutableArray alloc] init];
for (int i = 0; i < 3; i++)
{
//this string is autoreleased, you don't have call release on it.
//methods with the format [CLASS CLASSwithsomething] tend to be autorelease
NSString *s = [NSString stringWithFormat:#"%d", i);
[stagePickerArray addObject:s];
}
...
}
I think the only thing you are missing is a call to release in your dealloc method
-(void) dealloc
{
[stagepickerarray release]; //Do this for each of your arrays
[super dealloc];
}
The leaks tool will only tell you where yo allocated the objects that it thinks leaks. So, it's telling you, for instance, that
NSString* answer = [NSString stringWithFormat: ...
allocates an object that is never deallocated. Now, -stringWithFormat: gives you an object that you do not own and you don't seem to retain it anywhere. Therefore, you do not need to release it, so it can't be leaking by itself.
That means something else that you do own must be retaining it and you never release that something else. The prime suspect would appear to be stagePickerArray. Check that you are releasing stagePickerArray somewhere. If it's local to -viewDidLoad it must be released or autoreleased before the end of that method. If it's an instance variable, it must be released in the class's -dealloc method.
In Objective-C you need to take care of the retain count of allocated memory. If you don't need it -> release it.
Whenever you alloc an object, it will return an object with retain count = 1.
By using retain, the retain count gets incremented,
by using release, the retain count gets decremented.
Whenever the retain count is equals 0, the object will be destroyed.
So whenever you want to use the object somewhere else you need to retain it. So you make sure that the object is not deleted after the other 'person' (or whatever it used ;)) called release.
This was a very very very short description. Check the following guide
Memory Management Guide for iOS.
(Also you want to read something about ARC - Automatic Retain Counting - which is new in iOS5! ios5 best practice release retain
Quick one folks. Have a quick look at the code snippet below. I alloc the UILabel *textLabel but I don't release it (commented out). When the method ends, I lose reference to the pointer so it leaks.
Thing is that XCode Instruments doesn't find this leak and I would have thought that it is quite a straightforward case for it to find it. Its not reporting any leaks in my application and yet I found this one myself and its made me a little suspicious.
Am I over looking something or is Instruments rather poor at detecting leaks?
-(UITableViewCell*)newReadOnlyCellWithTitle:(NSString*)title andText:(NSString*)text {
UITableViewCell *cell=[[[UITableViewCell alloc] initWithStyle:UITableViewCellStyleDefault reuseIdentifier:nil] autorelease];
cell.textLabel.text=title;
cell.selectionStyle=UITableViewCellSelectionStyleNone;
cell.backgroundColor=[UIColor whiteColor];
cell.opaque=YES;
UILabel *textLabel=[[UILabel alloc] initWithFrame:CGRectMake(80, 11, 350, 24)];
textLabel.text=text;
textLabel.textColor=[UIColor lightTextColor];
textLabel.font=[UIFont fontWithName:STANDARD_FONT size:16.0];
textLabel.opaque=YES;
textLabel.backgroundColor=[UIColor clearColor];
[cell.contentView addSubview:textLabel];
//[textLabel release]; //<== without the release this should leak, yep?
return cell;
}
Edit: output from static analyizer...
Instrument's leak detection works by conservatively scanning memory, looking for pointers and building a graph of connections between allocations. If it finds any pointer(s) to an object in memory that can be reached from a global variable or stack variable, then that object cannot be considered leaked.
Instruments does not know the layout or context of a pointer. If you were to malloc(1024) and there happened to be a bunch of pointers in that [recycled] chunk of memory, those would count even though you will never treat those pointers as real references again.
So, no, Leaks can never be 100% accurate. As well, there are far more many ways to leak memory than an actual leak. If you had a global cache, like this:
NSMutableDictionary *myGlobalCache;
And you were to fill that cache but never prune it, that would be an effective memory leak and there is no way that it'll every show up in Instruments.
I wrote up an in depth discussion of Heapshot analysis, which is related and may be of interest.
Let me guess, -newReadOnlyCellWithTitle:andText: gets called by your -tableView:cellForRowAtIndexPath: method when cell == nil?
The thing is, those cells don't get released, they are cached and pulled back by -dequeueReusableCellWithIdentifier:. Table view uses same cells over and over again thus reducing the need to constantly allocate and release memory.
To make sure this is the case, you can try switching to different view controller, then simulating memory warning in simulator. This will force release on previous controller's view and as a result on cells in discussion. You should notice a leak in Instruments.
I think u will see it only when there are no valid references.
So, in your case, there is still a valid reference that goes something like, (probably) Screen -> table -> cell -> label. Looks like there is a possibility that your screen, the primary view, is still in memory, and hence still a valid reference.
You might want to check the objects that are allocated and then not released and still stay in the memory as valid objects. Checking for leaks isn't always enough.
Leaks will only find objects that are not referred to by any other object. For example if you create two objects in a method that refer to each other but no one else refers to either of them they are both technically leaked objects but leaks can see them since they each have an object that refers to them. Static analyzer is pretty good at finding these kinds of things though so the two in concert should kill almost any leak.
Your missing that there are 2 pointers to the object
Locally you have
UILabel *textLabel=[[UILabel alloc] initWithFrame:CGRectMake(80, 11, 350, 24)];
And within the windowing system which will manage it for you (it is my understanding it is not making a copy, but taking ownership)
[cell.contentView addSubview:textLabel];
So release is unnecessary, and with my limited knowledge even bad.
//[textLabel release]; //<== without the release this should leak, yep?
I am having a strange issue with tableview showing (null) when another ViewController comes and goes. I traced it down to the place that it reads the data to display to the tableview and found that it read null for data. To debug this, I need to understand how memory management really works for iPhone. I used the following code in my application delegate to create an array of objects. This was from a sample code which I modified for my app. My question is when one creates an array like this, would the [tempArray release] actually release the array that was just created?
//Initialize the hardware array.
NSMutableArray *tempArray = [[NSMutableArray alloc] init];
self.hardwareArray = tempArray;
[tempArray release];
This type of code seems to be very common for iPhone. The tutorials keeps saying for example, now that you passed the data to the controller you can release it. But my C and C++ experience tells me, if you release the memory it would be gone, and if its address is sitting someplace on the stack, its no good. Is the data being copied to another place rather than address being passed?
Also when releasing memory in a set routine like the one below, should I be checking to see if the modleName is not nil before releasing it. What happens if one released modelName that is already nil?
-(void) setModelName:(NSString *)newValue {
self.isDirty = YES;
[modelName release];
modelName = [newValue copy];
}
My question is when one creates an array like this, would the [tempArray release] actually release the array that was just created?
If the #property of the object is assign it would get released. But usually you use retain or copy as #property. And if the property is retain, the setter you call with self.hardwareArray = ... (ie [self setHardwareArray:...]) will retain the object. If the property is copy it will create a new object that is retained already.
Also when releasing memory in a set routine like the one below, should I be checking to see if the modleName is not nil before releasing it. What happens is one released modelName that is already nil?
In objective-c messages sent to nil are perfectly legal. No need to worry, no need to check for nil.
If you release an instance variable outside of dealloc you should set it to nil after releasing it.
release doesn't do what you think - it isn't equivalent to free() or delete and doesn't cause the memory to be gone.
It really is important to fully understand the few simple memory management rules before you do any coding. It's not difficult, the best resources are the Apple Developer Guides, like this one
http://developer.apple.com/library/ios/#documentation/Cocoa/Conceptual/MemoryMgmt/MemoryMgmt.html%23//apple_ref/doc/uid/10000011i
Running a static analysis with clang in both XCode 3.2 and Nikita Zhuk's Analysis Tool I've often come across this pair of warnings:
Method returns an Objective-C object with a +0 retain count (non-owning reference)
Incorrect decrement of the reference count of an object is not owned at this point by the caller
An example of code that may provoke this warning:
UIButton* button = [[UIButton buttonWithType:UIButtonTypeCustom] initWithFrame: CGRectMake(x, y, width, height)];
return button;
I assumed that buttons created this way are autoreleased, just like any other NSObject created with a convenience factory method. So I return it, and the caller can decide whether to retain it or not. What's the problem with that?
Am I obliged to retain and autorelease the object before returning it?
And most importantly, could whatever this warning is warning against ever be the cause of scary release-related crashes?
I realize now that this seems to only occur with UIButtons. Is it due to it being a class cluster?
EDIT: The snipped below shows a minimal case where clang issues these warnings (with the warnings in bold). Both warnings are flagged on the statement creating the object (the buttonWithType: message).
-(UIButton*) ztupidTezt:(UIImage*) img
{
UIButton* bt = [[UIButton buttonWithType:UIButtonTypeCustom]initWithFrame:
1
Method returns an Objective-C object with a +0 retain count (non-owning reference)
2
Incorrect decrement of the reference count of an object is not owned at this point by the caller
CGRectMake(0.0f, 0.0f, img.size.width, img.size.height)];
bt setImage:img forState:UIControlStateNormal];
return bt;
}
Well.... that code makes no sense.
buttonWithType: returns an instance of a UIButton that is already initialized. You shouldn't be calling -initWithFrame: on it.
Call setFrame:.
The bad code is confusing the analyzer.
Secondly, why bother with a third party tool to do the analysis. If you are using Xcode 3.2 on Snow Leopard (you should be -- it is a vastly better version of Xcode than the last release on Leopard), you can just "build and analyze". All of the analysis results will be presented inline with your code quite nicely.
The cause is most likely the use of sending both a buttonWithType: and initWithFrame: message. init* methods perform tasks that should only be done once for a given object. Class methods that create objects also initialize them. The result of your code is repeated initialization. Instead, send the buttonWithType message, then assign to the frame property.
Does the method name that has this code within it have "new" as a prefix? The Clang Static Analyzer follows standard Cocoa naming conventions and assumes that a -newSomething method will return an instance with a retain count of 1. If it's seeing an autoreleased object being returned from such a method, it might present the warning you're seeing.
Old, old, question. I am having the same problem. I think the existing answers successfully explained why the code is wrong, and why the analyzer says "+0 retain count". However it doesn't look like anybody explained why the analyzer says the code is decrementing the retain count. I think i figured out why. It's because of how init methods are allowed to return a different object than you sent the message to. They would release the original object, alloc a new one, and return it. The analyzer is assuming that any init method could do such a thing, although the init method in this example might not.