This question already has answers here:
When to use -retainCount?
(11 answers)
Closed 9 years ago.
Here is my code.
- (void)viewDidLoad{
[super viewDidLoad];
UIView *view = [[UIView alloc] initWithFrame:CGRectMake(100, 100, 240, 280)];
[view setTag:101];
UIImageView *imgView = [[UIImageView alloc] initWithFrame:CGRectMake(10, 10, 220, 260)];
[view setBackgroundColor:[UIColor redColor]];
[view addSubview:imgView];
[self.view addSubview:view];
[self getimageFromView:view];
[view release];
[imgView release];
[self getimageFromView];
}
-(void)getimageFromView:(UIView *)view{
for (UIView *view123 in [view subviews]) {
if ([view123 isKindOfClass:[UIImageView class]]) {
UIImageView *imgView = (UIImageView *)view123;
imgView.image = [UIImage imageNamed:#"img.png"];
NSLog(#"retain cnt 1 = %d",[imgView retainCount]);
}
}
}
-(void)getimageFromView{
for (UIView *view in [self.view subviews]) {
if (view.tag == 101) {
for (UIView *view123 in [view subviews]) {
if ([view123 isKindOfClass:[UIImageView class]]) {
UIImageView *imgView = (UIImageView *)view123;
imgView.image = [UIImage imageNamed:#"img.png"];
NSLog(#"retain cnt 2 = %d",[imgView retainCount]);
}
}
}
}
}
nslog look like following
retain cnt 1 = 3
retain cnt 2 = 2
Now my questions
1) Why UIImageView's object retain count is displayed like this ?
2) Is that correct count ?
3) If yes how can i send the release message till it become 0 ?
4) Can I do like this ? Is this proper way ?
for(int i=0;i<[imgView retainCount];i++){
[imageView release];
}
I have number of views like this and have to do operation on UIImageView as displayed.
Also I am getting memory warning and my app getting crash.
1) It is displayed like that because that is the retain count.
2) Yes
3) Absolutely not
4) See the answer to 3
The documentation for retainCount starts with the statement "Do not use this method." While this is a tad harsh the warning is there because it is very hard to interpret its result.
Answers to your questions:
1) Because that was its value at the time. A value of 3 means that at the instant was returned three objects had expressed ownership interest in the object, and that ownership interest was still outstanding but those objects may have already indicated they wish to cancel that interest by issuing an autorelease - this is one of the reasons retainCount should not be used.
2) The count is always correct at the time it is given, but it can be very hard to interpret.
3) You never, ever attempt to reduce the count to zero, that is completely at odds with the retain/release model and will cause havoc, or worse. The rule is if you create it (using alloc, or methods with create or copy in the names) or retain it then you have an ownership interest and must either hand that ownership to someone else or release/autorelease the object to cancel the ownership interest.
4) See 3, you never get here.
You should read the documentation on the retain/release model and understand ownership. Even under ARC, where retain/release is handled automatically for you, you need to understand ownership.
HTH
Related
I have a very strange error that I have never seen before. I have used this code hundreds of times before, but for some reason, the afterDelay is being completely disregarded.
My Code:
- (void)runIntro
{
if([self introDone])return;
UIImageView* back = [[UIImageView alloc] initWithFrame:[[AppDelegate shared] frame]];
[back setImage:[[AppDelegate shared] bgImage]];
[back setContentMode:UIViewContentModeScaleAspectFill];
[self.view addSubview:back];
UIImageView* target = [[UIImageView alloc] initWithFrame:CGRectMake(([[AppDelegate shared] width]-250)/2.0, ([[AppDelegate shared] height]-350)/2.0, 250, 72)];
[target setImage:[UIImage imageNamed:#"bpside.png"]];
[self.view addSubview:target];
Parabolic* wally = [[Parabolic alloc] initWithFrame:CGRectMake(([[AppDelegate shared] width]-200)/2.0, ([[AppDelegate shared] height]-314)/2.0+250, 200, 200)];
[wally setImage:[UIImage imageNamed:#"wallyside1.png"]];
[self.view addSubview:wally];
[self performSelector:#selector(moveWallyIntro:) withObject:wally afterDelay:1];
[self performSelector:#selector(moveWallyIntro:) withObject:wally afterDelay:2];
}
- (void)moveWallyIntro:(Parabolic*)wally
{
[wally runPBM:50 withDY:300 withDuration:0.5];
}
I know that all of the other code is working properly, but for whatever reason, the selectors are being performed immediately, which causes a major problem. This code is being run at the end of the viewWillAppear method. I have programmed this way many times before and this is the first time I've seen this happen.
UPDATE:
If I remove the code that adds the back UIImageView, this no longer is a problem. How can I add a giant image (3000x2000) in aspect fill format like this so it doesn't cause the program to hang? I'm using the same background image for iphone, ipad, etc., so it needs to be large for the retina iPads...
I've done the following thing:
buttonPlaylistView = [[UIButton alloc] initWithFrame:CGRectMake(self.view.frame.size.width *(urlList.count+1), 0, self.view.frame.size.width, self.view.frame.size.height)];
buttonPlaylistView.tag = 0;
UITapGestureRecognizer *doubleTap3 = [[UITapGestureRecognizer alloc] initWithTarget:self action:#selector(handleDoubleTap:)];
[doubleTap3 setNumberOfTapsRequired:2];
[buttonPlaylistView addGestureRecognizer:doubleTap3];
[doubleTap3 release];
-(void) handleDoubleTap:(UITapGestureRecognizer *)sender{
if(sender.state == UIGestureRecognizerStateEnded)
int x = [sender tag];
return;
}
But I get SIGAGRT at this line: int x = [sender tag]; saying:
[UITapGestureRecognizer tag]: unrecognized selector sent to instance 0x61280b0
*** Terminating app due to uncaught exception 'NSInvalidArgumentException', reason: '-[UITapGestureRecognizer tag]: unrecognized selector sent to instance 0x61280b0'
NOW:What's the problem and what's the solution for this?Thanks
-(void) handleDoubleTap:(UITapGestureRecognizer *)sender
{
if(sender.state == UIGestureRecognizerStateEnded)
{
int x = [sender.view tag];
}
}
Will fix the issue.
An UITapGestureRecognizer does not have a property named tag - as you see, the sender you get is NOT the button. You have to access the buttonPlayListView directly, like
int x = [buttonPlayListView tag];
or otherwise remember which button you want to access.
Even though I'm quite sure that you're going about this the wrong way, adding a double tap gesture recognizer to a UIButton, there is a way you can still perform the task you require that shouldn't be too much work for you.
You've made the comment
and how could I remember if I create let say 100 buttons
to one of the answers, the one which highlights what the issue is that's causing your SIGBART. UIGestureRecognizer does not have a tag property.
Here's what you could do, is to iterate through all the subviews of your [self view] and find the one that has the same UIGestureRecognizer, it's not the prettiest solution, and the more subview's you have the longer the loop will take. But it'll do what you seem to be looking for, so if you're adding .
In your handleDoubleTap function you could do the following
-(void) handleDoubleTap:(UITapGestureRecognizer *)sender
{
if(sender.state == UIGestureRecognizerStateEnded)
{
int iButtonTag = -1 //This is important later to escape the below for loop as we don't need to needlessly go around in circles
for(UIView* psubView in [[self view] subviews])
{
if( [psubView isKindOfClass:[UIButton class]] )
{
UIButton* pButton = (UIButton*)psubView;
for(UIGestureRecognizer* pGesture in [pButton gestureRecognizers] )
{
if( pGesture == sender )//this is the button we're after
{
iButtonTag = [pButton tag];
break;
}
}
if( iButton != -1 )//found what we came for
{
break;
}
}
}
//do what ever it was you needed to do now that you have the views tag, or you could have kept a reference to the button etc.
}
}
That should solve your problem. Alternatively if you're going to be adding buttons to subviews of subviews it would be better to keep track of your UIButtons in an NSMutableArray , you would do this by creating a class property (or member variable) and adding the buttons to this using the 'addObject:' function of NSMutableArray. Then instead of the line
for(UIView* psubView in [[self view] subviews])
above you could exchange that for
for( UIButton* pButton in m_pMutableButtonArray )
where "m_pMutableButtonArray" is the variable name you gave to your NSMutableArray you were storing the UIButtons in. This also means you would do away with the following if isKindOfClass test on the following line.
That should fix your problem.
Why are you putting a UITapGestureRecognizer in a button? The button already handles that for you and will send you a callback, you can add a target to a button using this UIControl method
- (void)addTarget:(id)target action:(SEL)action forControlEvents:(UIControlEvents)controlEvents
I have a problem with removing a subview, or more precisely with checking if it is still there after having deleted it.
My app first adds a subview to self.view.
[self.view addSubview:tabsClippedView];
Then it adds another subview to this subview (to which it adds several buttons as subviews, but I guess this is unimportant in this context):
[tabsClippedView addSubview:tabsView];
Finally, I have a method which allows the tabsView to be deleted and then created again. I need to do this so as to update the number of buttons in that tabsView (as the user can delete buttons). The method looks basically like this:
[self.tabsView removeFromSuperview];
After that I call a method called showTabs (which I already called in the very beginning of the app in order to add the subViews). This is where it all becomes problematic and where my app crashes (I get no error in the debug console, so I don't really know what the issue is...):
if ([tabsClippedView isDescendantOfView:self.view]) {
NSLog(#"There is already a tabsClippedView.");
} else {
NSLog(#"There is no tabsClippedView. I'll add one...");
[self initTabsClippedView];
}
This is where the app crashes: when trying to assess if tabsView isDescendantOfView (I don't get any of the following logs):
if ([tabsView isDescendantOfView:tabsClippedView]) {
NSLog(#"There is already a tabsView");
} else {
NSLog(#"There is no tabsView for the buttons. I'll add one including buttons.");
[self initTabs];
}
I'd be grateful for any suggestions where the problem could be.
EDIT:
These are the methods to set up my views:
-(void) initTabsClippedView { // sets up tabsClippedView
NSLog(#"initTabsClippedView method started...");
CGRect tabsClippedFrame = CGRectMake(258,30,70,81*6);
tabsClippedView = [[[UIView alloc] initWithFrame:tabsClippedFrame] autorelease];
tabsClippedView.clipsToBounds = true;
[self.view addSubview:tabsClippedView];
NSLog(#"initTabsClippedView method ended.");
}
-(void) initTabs {
NSLog(#"initTabs started. Adding buttons to tabsClippedView...");
CGRect tabsFrame = CGRectMake(-30,0,50,480);
tabsView = [[[UIView alloc] initWithFrame:tabsFrame] autorelease];
[tabsClippedView addSubview:tabsView];
// sets up buttons in tabsClippedView
And this is where I delete the tabsClippedView (triggered by a button found in tabsClippedView):
-(void)tabDelete:(id)sender
{
UIButton *button = (UIButton *)sender;
[UIView animateWithDuration:0.75
delay:0
options:UIViewAnimationOptionCurveEaseIn
animations:^{
button.transform = CGAffineTransformMakeTranslation(-30, 0);
}
completion:^(BOOL finished){
[UIView animateWithDuration:0
delay:0
options:UIViewAnimationOptionCurveEaseIn
animations:^{
[self.tabsView removeFromSuperview];
//...
}
completion:^(BOOL finished){
NSLog(#"tabsView removed from Superview. Objects Deleted.");
[self showTabs];
NSLog(#"TabDelete finished. Button removed and tabsView updated accordingly.");
}
];
}];
And this is the showTabs method which was already called when I started the app:
-(void)showTabs {
NSLog(#"showTabs started...");
currentView = #"Tabs";
if ([tabsClippedView isDescendantOfView:self.view]) {
NSLog(#"There is already a tabsClippedView.");
} else {
NSLog(#"There is no tabsClippedView. I'll add one...");
[self initTabsClippedView];
}
if ([tabsView isDescendantOfView:tabsClippedView]) {
NSLog(#"There is already a tabsView");
} else {
NSLog(#"There is no tabsView for the buttons. I'll add one including buttons.");
[self initTabs];
}
Is it possible that you are getting EXC_BAD_ACCESS? Is it possible that the app is crashing because tabsView is deallocated when you send isDescendantOfView: to it. If you run with breakpoints enabled it should tell you the reason for the crash. If it is an EXC_BAD_ACCESS problem you should try NSZombie.
To activate NSZombie do the following:
Get info of the executable.
Go to the arguments tab.
In the "Variables to be set in the environment:" section add:
Name: NSZombieEnabled
Value: YES
Then run your app as usual and when it crashes it should tell you which deallocated object received what message.
EDIT: Just saw your edit. I think I nailed it. You're autoreleasing the views when you create them, so when they are removed from their superviews they are no longer retained and thus deallocated. You're app crashes because you're trying to run methods on deallocated views.
EDIT 2: Thought I should tell you that there is a better solution than the one posted by Praveen S.
Change your code as follows:
[tabsClippedView release];
tabsClippedView = [[UIView alloc] initWithFrame:tabsClippedFrame];
and
[tabsView release];
tabsView = [[UIView alloc] initWithFrame:tabsFrame];
The above code does the same thing as the code posted by Praveen S, but without the autorelease. An autorelease is more expensive than a regular release and should only be used when needed and in this case it isn't.
Rather than releasing before you allocate a new view you probably want to release the view when you're done with it:
[tabsView removeFromSuperview];
[tabsView release];
tabsView = nil;
or simply
[tabsView removeFromSuperview];
self.tabsView = nil;
and then instead of:
if ([tabsView isDescendantOfView:tabsClippedView]) ...
you can use:
if (tabsView) ...
As you might have noticed, there really is no need for you to retain the view. You could just as well do the following:
tabsView = [[UIView alloc] initWithFrame:tabsFrame];
[tabsClippedView addSubview:tabsView]; // This retains tabsView
[tabsView release];
and then to remove the view you would use:
[tabsView removeFromSuperview]; // This will release the tabsView
tabsView = nil;
Also remember to set the views to nil in viewDidUnload.
EDIT 3: Why self made such a difference:
What you need to understand is how properties and reference counting works. There are books and such you could read about it. I'm sure Google can provide you with some good references as well.
The difference between
self.tabsView = [[UIView alloc] initWithFrame:frame];
and
tabsView = [[UIView alloc] initWithFrame:frame];
is that self.tabsView is accessing the properties setter, while tabsView is accessing the instance variable directly.
A nonatomic, retain property's implementation looks something like the following:
- (void)setTabsView:(UIView *)view
{
if (view != tabsView) {
[tabsView release];
tabsView = [view retain];
}
}
So the property is taking care of the memory management for you. In my solution I take care of the memory management myself and thus I don't need the property to do it for me, so I don't use it.
I hope this explains why self made such a difference.
Change your code as follows:
self.tabsClippedView = [[[UIView alloc] initWithFrame:tabsClippedFrame] autorelease];
and
self.tabsView = [[[UIView alloc] initWithFrame:tabsFrame] autorelease];
I'm trying to scroll through images being downloaded from a users online album (like in the facebook iphone app) since i can't load all images into memory, i'm loading 3 at a time (prev,current & next). then removing image(prev-1) & image (next +1) from the uiscroller subviews.
my logic works fine in the simulator but fails in the device with this error:
[CALayer retain]: message sent to deallocated instance
what could be the problem
below is my code sample
- (void)scrollViewDidEndDecelerating:(UIScrollView *)_scrollView
{
pageControlIsChangingPage = NO;
CGFloat pageWidth = _scrollView.frame.size.width;
int page = floor((_scrollView.contentOffset.x - pageWidth / 2) / pageWidth) + 1;
if (page>1 && page<=(pageControl.numberOfPages-3)) {
[self removeThisView:(page-2)];
[self removeThisView:(page+2)];
}
if (page>0) {
NSLog(#"<< PREVIOUS");
[self showPhoto:(page-1)];
}
[self showPhoto:page];
if (page<(pageControl.numberOfPages-1)) {
//NSLog(#"NEXT >>");
[self showPhoto:page+1];
NSLog(#"FINISHED LOADING NEXT >>");
}
}
-(void) showPhoto:(NSInteger)index
{
CGFloat cx = scrollView.frame.size.width*index;
CGFloat cy = 40;
CGRect rect=CGRectMake( 0, 0,320, 480);
rect.origin.x = cx;
rect.origin.y = cy;
AsyncImageView* asyncImage = [[AsyncImageView alloc] initWithFrame:rect];
asyncImage.tag = 999;
NSURL *url = [NSURL URLWithString:[pics objectAtIndex:index]];
[asyncImage loadImageFromURL:url place:CGRectMake(150, 190, 30, 30) member:memberid isSlide:#"Yes" picId:[picsIds objectAtIndex:index]];
[scrollView addSubview:asyncImage];
[asyncImage release];
}
- (void) removeThisView:(NSInteger)index
{
if (index<[[scrollView subviews] count] && [[scrollView subviews] objectAtIndex:index]!=nil) {
if ([[[scrollView subviews] objectAtIndex:index] isKindOfClass:[AsyncImageView class]] || [[[scrollView subviews] objectAtIndex:index] isKindOfClass:[UIImageView class]]) {
[[[scrollView subviews] objectAtIndex:index] removeFromSuperview];
}
}
}
For the record it works OK in the simulator, but not the iphone device itself.
any ideas will be appreciated.
cheers,
fred.
I guess your AsyncImageView class uses asynchronous connections to load the image (which is a good thing), so that means that when you add it as a subview of your scroll view, it might be "removedFromSuperview" before the image loads completely. Check in the implementation of your class if the delegates are properly nil'd when it's been dealloc'd, and also that any connections are cancelled and so on.
Pay attention to the fact that a class should never retain its delegate, an "assign" kind of property is perfect for these situations.
Some tips:
Instead of alloc/init'ing and release'ing so many instances while you scroll, try to keep three ivars in your class, and just change the "frame" property of the instances as you scroll. You will avoid lots of allocation and releases if the user scrolls fast, and that's a good thing on an iPhone.
Avoid direct ivar accesses like that in your code; use properties, there are many advantages to them beyond the simple "correctness" thing (KVO being one of them).
I'm trying to create an activity indicator in iPhone app. The problem is that I cannot get it to appear before the actual task i want it to diplay during is already done. Is there something funky about the order in which the iPhone does stuff?
Here is my problematic code (in my app delegate):
-(BOOL)showProgressView: (NSString *) message {
self.progress = [[UIView alloc] initWithFrame:window.frame];
UIImageView *img = [[UIImageView alloc] initWithImage:[UIImage imageNamed:#"squircle.png"]];
[img setAlpha:0.5];
[img setFrame:CGRectMake(94, 173, 133, 133)];
UIActivityIndicatorView *spinner = [[UIActivityIndicatorView alloc] initWithFrame:CGRectMake(51.5, 51.5, 30, 30)];
spinner.activityIndicatorViewStyle = UIActivityIndicatorViewStyleWhiteLarge;
[img addSubview:spinner];
[self.progress addSubview:img];
[spinner startAnimating];
[img release];
[spinner release];
[window addSubview:self.progress];
return YES;
}
I then call this code like this:
if ([appDelegate showProgressView:#"Loading..:"])
{
//My actual code loads data and stuff here but that is not important
//drawCtrl is a UIViewController subclass that is instantiated here
UINavigationController *navController = [appDelegate navigationController];
[navController pushViewController:drawCtrl animated:YES];
[drawCtrl release];
}
The problem is that my activity indicator does not appear until the new view controller is pushed onto the navController's stack. Can I control this in some way?
Thanks in advance!
-Mats
You need to call your loading method periodically via a timer. UI changes require a trip through the event loop to be seen. I have done this usually in a timer I set up in the AppDelegate, it calls the "methodThatTakesSomeTime" every n seconds, the "methodThatTakesSomeTime" method does some work for a specified time slice and then exits which is usually about the same as the timer trigger time. This gives the event loop time to refresh the UI and also give your method time to do its stuff.
Takes some fiddling with keeping the state of the "methodThatTakesSomeTime" so it can continue on its way, but that is what it takes.