I'm having a view controller which holds a scroll view, a content view and a CATiledLayer as sublayer in the content view.
If I remove my controllers view from its superview while the CATiledLayer is still busy (rendering a PDF page for instance), I get the weirdest crashes and null references. It seems like CATiledLayer is not happy if you disturbe it. Is there a way I can abort what it is currently doing?
Am I right that the controller you are removing is the delegate of the CATiledLayer?
Then you have to set the CATiledLayer's delegate to nil when you remove your controller.
May be related link (my own question XD): CATiledLayer drawLayer:inContext: crashing when the view is deallocated while the image to draw is being retrieved
Related
I have a following situation (testing on iPad, iOS 5.1):
There is a UIPopoverController with UINavigationController inside and custom popover background view (subclass of UIPopoverBackgroundView).
There is a generic UIViewController (let's call it VC1) as the root VC in the Navigation Controller.
I push another UIViewController (VC2) with UITableView on the Navigation Controller stack.
Effect:
Table scrolling is choppy (looks like 10-15 fps). For testing purposes I use a simplest possible UITableView, without images etc. so it's NOT caused by bad UITableView implementation.
Scrolling is not choppy if the VC2 is the root view controller of
Navigation Controller, even with the custom Popover background.
It's also not choppy if pushed as the second VC but I don't use custom bg view class for UIPopoverController.
I log each of the overriden methods inside my UIPopoverBackgroundView subclass, and they aren't called constantly or anything, which could theoretically cause performance hit. I'm going to debug the problem further but maybe someone has already solved it before?
Or maybe someone has good suggestions on how to find the culprit? I tried looking into time profiler for offending function calls, but I didn't find much there...
I have implemented a UIWebView object into a custom UIView (TSAlertView). My UIWebView sits in the middle of the UIView, as a subview.
When I rotate the iOS device screen, the UIWebView object redraws right on top of the old one, without deleting the old one. All the other elements of this custom UIView are destroyed and redrawn when the screen is rotated.
I imagine I must have missed some sort of procedure for deallocating resources or removing the UIWebView from view. I have tried adding 'autorelease' to the declaration of the UIWebView, but to no avail. I wonder if this is a common symptom of a simple oversight I have made in the way the UIWebView is created?
This seems to be a simple case of not telling the UIWebView object it should destroy itself before it is redrawn on each rotate - but I don't know how I can go about this...
Any wisdom gratefully received.
Ideally while rotating you should not create the views again. Instead you can rearrange the frame in delegates appropriately. In your case you might be creating a UIWebView again, which should be avoided. Instead of that, you can keep it as a class level param(Declare in .h file) and adjust the frame in delegate methods for rotation. Also please note that release/autorelease wont remove your view from superview. You need to call removeFromSuperview method to achieve that.
Check this for view resizing and layout handling
I have a UINavigationController within one of the tabs of a UITabBarController.
I now present a new view controller (let’s call it Steve) over the whole app (using presentViewController:animated:completion:).
Then, I simulate low memory.
After dismissing Steve (using dismissViewControllerAnimated:completion:), I can now see that the UINavigationController’s view is gone; within the tab; only an empty white area is seen!
Why is this? I have tried calling the view methods on all imaginable controllers upon Steve’s dismissal, but the contents of the tab still stay empty (white).
The strange thing is this: If I click on another tab, and click back on the original tab, the contents (the navigation controller) shows just fine again. Is the tab bar controller doing something special to force the view to display?
UPDATE: I was able to “fix” my issue with this terrible code, just before dismissing Steve:
[[[[[self tabBarController] view] subviews] objectAtIndex:0]
addSubview:[[self navigationController] view]];
What this does is that it finds the subview of the tab bar controller which is not the tab bar (i.e. the top view), and then adds the navigation controller’s view to be its subview.
This is of course terrible, because it makes internal assumptions about the subview structure of the tab bar controller’s view.
If someone has any better solutions, please let me know about them.
When your app receives a memory warning, one of the first thing it will do is drop the view hierarchies of any view controllers that have loaded views but are not currently visible (like your UINavigationController). Most likely, whatever view controller is at the top of your navigation stack is having its views dropped but not reloading them when it reappears.
Always put your view construction code in -loadView or -viewDidLoad and not in -init. That way, the view controller will reconstruct your view if it's been dropped due to a memory warning.
(P.S.: The reason your hack works is the bit where you call [[self navigationController] view], which in turn calls -loadView on the top VC in the stack, forcing it to reconstruct its view.)
I downloaded the free iPad application "SketchBook Express" by "Autodesk".
In it, they have a drawing surface, and (what I believe is) a UINavigationController. What interests me is that the navigation controller bar rotates with the iPad, but the content of the drawing does not. So if I put the iPad from portrait to landscape, the nav bar moves to the new top, but the sketch I've drawn is now on its side. I would like to implement something similar for my program, but I have been unable to.
I've played with shouldAutorotateToInterfaceOrientation: but so far I've just figured out it is only called once. I made a subclass of UINavigationController and had that return YES, and had my subview controller return NO. That did not do the trick.
Is there a simple way to do this?
What rotates is probably contained by a UIViewController (or UINavigationController)'s view.
Other things (that don't rotate) are probably contained by a view that is not contained by any UIViewController (or a controller that returns NO in shouldAutorotateToInterfaceOrientation method).
This means that you could try setting your drawing view as a subview of the keyWindow.
Other things could be in the view controller as usual.
Is not I have tried, this is just an idea. And I would be interested in knowing if this worked. :)
I am trying to detect touches, but the touchesBegan method is not being called.
In my ViewController, I have added the touchesBegan method. My Nib files owner is set to the correct V.C. The Nib itself consists of the view, with a scroll view and a tab bar. Nested in the scroll view is an image view, which has user interaction enabled. What is precluding touches from being registered, or preventing my implementation of touchesBegan from being called?
I've scoured the Internet and Apple docs, and I can't see what I am doing wrong. Also, I'm not really sure what code I can post here to help with my query. Thanks.
Okay, after a lot more reading, I've now got a scrollview and a imageview, both of which are created programatically. The imageview is a sub view of the scrollview, and scrollview has been subclassed so that the touches ended method can decide whether it was a single touch, in which case call the touches ended method from the view controller, otherwise call its supers method. This works just fine, however, why is it that this cannot be done without subclassing scrollview? Is it my lack of understanding of how scrollview works, or is it just a limitation of it?