i want to create a basic push view scenario. I have the following code but it does not work. Nothing happens. Could someone tell me why?
testController *screen2 = [[testController alloc] initWithNibName:nil bundle:nil];
[self.navigationController pushViewController:screen2 animated:YES];
[screen2 release];
You should try to do an NSLog to see if screen2 is nil. Chances are it is not being appropriately loaded from a corresponding nib for some reason. As an aside, I'd highly recommend sticking to the convention of capitalizing class names. Speaking of which, did you maybe call the nib file TestController.nib? (That would cause the problem.)
Related
It seems that UIPageViewController is holding the initial content view controller forever.
For example:
DataViewController *startingViewController = [self.modelController viewControllerAtIndex:0 storyboard:self.storyboard];
NSArray *viewControllers = #[startingViewController];
[self.pageViewController setViewControllers:viewControllers direction:UIPageViewControllerNavigationDirectionForward animated:NO completion:NULL];
self.pageViewController.dataSource = self.modelController;
The startingViewController is never released until the pageViewController itself it released.
To reproduce this bug, just create a new project in XCode using the Page-Based Application template. And add 3 lines of code into DataViewController.m
#property NSInteger debugIndex; // file scope
NSLog(#"DataViewController[%d] created", self.debugIndex); // in viewDidLoad
NSLog(#"DataViewController[%d] dealloc", self.debugIndex); // in dealloc
And when you scroll the demo App in vertical orientation, you'll get logs like this:
DataViewController[0] created
DataViewController[1] created
DataViewController[2] created
DataViewController[1] dealloc
DataViewController[3] created
DataViewController[2] dealloc
DataViewController[4] created
DataViewController[3] dealloc
DataViewController[5] created
DataViewController[4] dealloc
DataViewController[6] created
DataViewController[5] dealloc
DataViewController[0] is never deallocated.
Any ideas about this?
Thanks!
Are you using transitionStyle UIPageViewControllerTransitionStyleScroll? I encountered the same or a similar problem which seemed to disappear when using page curl animations instead.
The problem was compounded for me because I was allowing a UISliderBar to set the position in the content. So on change of the UISliderBar, I was calling setViewControllers:direction:animated:completion: which caused more and more view controller references to get "stuck" in my UIPageViewController.
I am also using ARC. I have not found an acceptable way to force the UIPageViewController to let go of the extra view controller references. I will probably either end up using the page curl transition or implementing my own UIPageViewController equivalent using a UIScrollView with paging enabled so I can manage my own view controller cache instead of relying on UIPageViewController's broken view controller management.
I'm not sure you still got the problem, but I had the same problem and I found the solution.
I don't know the reason, but it works.
I'm setting the first viewController right after addSubview, rather than before addChlidViewController.
-(void)settingPageViewController{
if (!self.pageViewController) {
self.pageViewController = [[UIPageViewController alloc]initWithTransitionStyle:UIPageViewControllerTransitionStyleScroll navigationOrientation:UIPageViewControllerNavigationOrientationHorizontal options:nil];
self.pageViewController.delegate = self;
self.pageViewController.dataSource = self;
[self addChildViewController:self.pageViewController];
[self.pageViewController didMoveToParentViewController:self];
[self.containerView addSubview:self.pageViewController.view];
[self.pageViewController setViewControllers:#[[self viewcontrollerAtIndex:0]] direction:UIPageViewControllerNavigationDirectionForward animated:NO completion:nil];
}
}
and the first viewController will dealloc in the right time.
also, I found if call
[self.pageViewController setViewControllers:#[[self viewcontrollerAtIndex:0]] direction:UIPageViewControllerNavigationDirectionForward animated:NO completion:^(BOOL finished){
NSLog(#"finished : %d",finished);
}];
before addSubView and the completion block will not call.
and I reckon this block is the reason why the first viewController didn't dealloc.
I'll go find out why it didn't callback, and improve the answer~
cheers
After a few attempts to figure out what was happening on a similar issue, I noticed that in my project there were 2 reasons that caused a retain problem and resulted in having a UIPageViewController being forever retained.
1) there was a circular reference between the UIPageViewController and the UIViewcontroller that was presented (this was fixed by changing the properties to weak from strong in both classes)
2) and the main fix consisted in changing
[self setViewControllers:#[initialDetailsViewController] direction:UIPageViewControllerNavigationDirectionForward animated:YES completion:nil];
to
__weak __typeof__(self) weakSelf = self;
[weakSelf setViewControllers:#[initialDetailsViewController] direction:UIPageViewControllerNavigationDirectionForward animated:YES completion:nil];
I hope this helps someone
same problem here
i resolved it by keeping my initial viewController in the variable
and instead of creating the same vc on particular pageIndex i just reuse it
I had same problem and solved the following:
[startingViewController release]; where the end point of initialization.
then the first ViewController will be deallocated.
I started using CoreData in my application following Stanford CS193P lessons regarding the use of iOS 5's new class UIManagedDocument. The approach itself is quite straightforward but I can't understand how to deal with model modifications i keep making. This is how I instantiate my UIManagedDocument object (inside the appDelegate, so that every other class can use it):
if (!self.database) {
NSURL *url=[[[NSFileManager defaultManager] URLsForDirectory:NSDocumentDirectory inDomains:NSUserDomainMask] lastObject];
url = [url URLByAppendingPathComponent:#"AppName"];
UIManagedDocument *doc = [[UIManagedDocument alloc] initWithFileURL:url];
NSDictionary *options = [NSDictionary dictionaryWithObjectsAndKeys:
[NSNumber numberWithBool:YES], NSMigratePersistentStoresAutomaticallyOption,
[NSNumber numberWithBool:YES], NSInferMappingModelAutomaticallyOption, nil];
doc.persistentStoreOptions = options;
self.database=doc;
[doc release];
}
The issue I have is that every time I change even a little bit of my .xcdatamodel, I am unable to get all the content previously stored in the document as well as to create any new instance. As a matter of fact doing this generates the following exception:
*** Terminating app due to uncaught exception 'NSInternalInconsistencyException', reason: 'This NSPersistentStoreCoordinator has no persistent stores. It cannot perform a save operation.'
I thought setting the "options" property of the managed document would have solved the problem, but apparently this isn't enough.
Anyone can help? Couldn't' find other questions that actually fit my precise needs.
Before you modify your Core Data model, you should "Add Model Version".
1.
Select the original model file. (e.g. YourProject.xcdatamodel)
2.
"Editor" -> "Add Model Version...". Then add a new model version (e.g. 2.0)
3.
You will get a new model file. (e.g. YourProject 2.0.xcdatamodel). Modify it.
4.
Change the current model version. Select the top .xcdtatmodel file -> "View" -> "Utilities" -> "Show File Inspector". Find the tag "Versioned Core Data Model" and choose the right version you want to modify.
It also disturbed me for a long long time. Hope this way can help you ^^
There is a really simple solution to your problem. Simply erase the app from the simulator or device (touch the app icon for a few seconds, and touch the cross that appears when the app icons start wiggling). That way, Xcode updates your app's UIManagedDocument.
Make sure that you did not mistype NSDocumentDirectory as NSDocumentationDirectory.
NSURL *url = [[[NSFileManager defaultManager] URLsForDirectory:NSDocumentDirectory inDomains:NSUserDomainMask] lastObject];
NSDocumentDirectory is right!
I'm analyzing an image which takes some time and meanwhile I want to display a progress indicator. For this I'm using MBProgressHUD.
It almost works... I get this error: "Modifying layer that is being finalized". I guess it's due to the fact that I do pushViewController not in my main thread. Am I right? Any ideas on how to correct this issue?
My code:
- (IBAction)buttonReadSudoku:(id)sender
{
mbProgress=[[MBProgressHUD alloc] initWithView:self.view];
mbProgress.labelText=#"Läser Sudoku";
[self.view addSubview:mbProgress];
[mbProgress setDelegate:self];
[mbProgress showWhileExecuting:#selector(readSudoku) onTarget:self withObject:nil animated:YES];
}
- (void)readSudoku
{
UIImage *image = imageView.image;
image = [ImageHelpers scaleAndRotateImage:image];
NSMutableArray *numbers = [SudokuHelpers ReadSudokuFromImage:image];
sudokuDetailViewController = [[SudokuDetailViewController alloc] init];
[sudokuDetailViewController setNumbers:numbers];
[[self navigationController] pushViewController:sudokuDetailViewController animated:YES];
}
Define a new method to push your detail view controller and use -performSelectorOnMainThread:withObject:waitUntilDone: to perform it on the main thread. Don't try to make any UI changes from other threads.
All UI changes must be in the main thread, as you note. Rather than you off-main-thread method make any changes to the UI, send an NSNotification to the current viewController telling it to do the UI work.
This is an especially good route if you're crossing an MVC border, or if you already have a viewController that knows what to do so that writing a separate method results in duplicate code.
This is an update to my problem. I am receiving this warning now when the program aborts.
warning: Attempting to create USE_BLOCK_IN_FRAME variable with block that isn't in the frame.
I can't find much information on what this means.
This has me baffled. I get the EXC_BAD_ACCESS error. I have NSZombieEneabled (which helped with an earlier problem), but there is no call stack to trace.
I have some nearly identical code that is working with respect to another fetched result controller.
This seems to have something to do with the relationships between the job entity and its associated client entity. The relationship is [job entity] <<--> [client entity].
Initially, I see that the code works without error where the job entity corresponding to the selected row has no client entity linked through a relationship. So in the case where it fails, this points to a client entity, but when it doesn't fail, the pointer is nil.
When I encounter this problem, I start the application and go directly to the job picker view and select a cell. It's at that point that the problem occurs.
I did an experiment by starting the application and going to the client picker view first, knowing that a fetch would occur of all of the client entities. Then I went to the job picker view and selected a cell. The problem did not occur.
Since I am just trying to pass a pointer to a job entity that was already fetched, I don't understand what's happening.
By the way, the code was working fine before I switched to using NSFetchedResultsControllers. I like what they can do for me, but there are some dynamics going on here that I haven't figured out.
The logging is not showing me anything I understand toward resolving the problem.
- (void)tableView:(UITableView *)tableView didSelectRowAtIndexPath:(NSIndexPath *)indexPath
{
userState.selectedJob = [self.fetchedResultsController objectAtIndexPath:indexPath];
NSLog(#"\n\n(1 Pick) indexPath: %#\n",indexPath);
NSLog(#"\n\n(1 Pick) userState: %#\n",userState);
NSLog(#"\n\nnumber of Objects in job fetchresultscontroller = %d", [[fetchedResultsController fetchedObjects] count] );
NSLog(#"\n\n(1 Pick) selected job: %#\n",[self.fetchedResultsController objectAtIndexPath:indexPath]); // This line is causing the problem...
NSLog(#"\n\n(1 Pick) selected job: %#\n",userState.selectedJob); // Omitting the line above, this line fails
[self.navigationController pushViewController:userState.jobInfoTVC animated:YES];
}
The debug output is
2011-05-07 09:27:04.142 job1[6069:207]
(1 Pick) indexPath: <NSIndexPath 0x5952590> 2 indexes [0, 3]
2011-05-07 09:27:04.142 job1[6069:207]
(1 Pick) userState: <UserStateObject: 0x5919970>
2011-05-07 09:27:04.143 job1[6069:207]
number of Objects in job fetchresultscontroller = 4
(gdb)
The final code should be as simple as this, which led me to all of the logging:
- (void)tableView:(UITableView *)tableView didSelectRowAtIndexPath:(NSIndexPath *)indexPath
{
userState.selectedJob = [self.fetchedResultsController objectAtIndexPath:indexPath]; // Original failure was at this line
[self.navigationController pushViewController:userState.jobInfoTVC animated:YES];
}
I use the singleton userState to keep track of what the user has done. So I keep last selectedJob and selectedClient entity pointers there. This has worked okay before I switched to NSFetchedResultsController.
I have also had a problem with Attempting to create USE_BLOCK_IN_FRAME variable with block that isn't in the frame. Although mine wasn't anything to do with NSFetchedResultsControllers.
Possibly Your Problem
I noticed you mention you are using a singleton, so maybe your problem is solved at this link:
http://npenkov.com/2011/08/01/solving-issues-like-warning-attempting-to-create-use_block_in_frame-variable-with-block-that-isnt-in-the-frame/
From the link:
Exactly in session manager I used the macro, before #synthesize – this was the problem, static definitions should not appear before synthesized methods. So if you have something like:
SYNTHESIZE_SINGLETON_FOR_CLASS(SessionManager)
#synthesize loggedUserId, ...
Just replace it with:
#synthesize loggedUserId, ...
SYNTHESIZE_SINGLETON_FOR_CLASS(SessionManager)
enter code here
My problem
The problem I had was to do with duplicating a variable declaration, in this case the variable inherited from NSManagedObject :
- (void)functionThatDoesSomething {
VariableInheritedFromNSMObj *variableA = nil;
for (VariableInheritedFromNSMObj *variableA in [self containerObject]) {
NSLog(#"a = %#\n", [variableA name]);
}
[variableA setName:#"StackOverflow"];
}
Moving the first line, where variableA is initialised to nil, to after the loop fixed the problem. In my production code I then changed the name of one of the variables.
Hope that helps you or someone else who comes across this problem. This error seems to manifest itself in many different ways.
Is the deployment target OS version set to a lower one than you have debugging support for in Xcode 4.2? When this happened to me, the deployment target had somehow changed to 4.1, but I only had the 4.3 simulator without the additional debugging support for 4.0-4.1.
There are ways to fix this:
Install the debug symbols for your lower OS version from the Xcode preferences. Go to downloads, components and install OS 4.0 - 4.1 Device Debugging Support (or earlier if you need it).
Set the deployment target to 4.3 or higher.
If you choose the former, it's probably worth ticking the box in preferences to download the updates automatically.
Up to iOS 3.2, I used this kind of code to load UIImageView image in background, and it worked fine...
Code:
- (void)decodeImageName:(NSString *)name
{
NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init];
UIImage *newImage = [UIImage imageNamed:name];
[myImageView setImage:newImage];
[pool release];
}
...
[self performSelectorInBackground:#selector(decodeImageName:) withObject:#"ID"]
... even if [UIImageView setImage:] was not thread-safe !
But since iOS 4, it doesn't work any more... Images appear on screen two seconds after setImage call. And if I do a [myImageView performSelectorOnMainThread:#selector(setImage:) withObject:newImage waitUntilDone:YES] instead of [myImageView setImage:newImage], images appear immediately but seem to be re-decoded again on-the-fly (ignoring the previous [UIImage imageNamed:] which should have already decoded the image data), causing a pause on my main thread... Even if documentation says The underlying image cache is shared among all threads..
Any thought ?
Don’t do it in the background! It’s not thread-safe. Since an UIImageView is also an NSObject, I think that using -[performSelectorOnMainThread:withObject:waitUntilDone:] on it might work, like:
[myImageView performSelectorOnMainThread:#selector(setImage:) withObject:newImage waitUntilDone:NO];
And it’s UIImage which is newly made thread-safe. UIImageView is still not thread-safe.
performSelectorInBackground: runs a selector in a background thread. Yet setImage: is a UI function. UI functions should only be run on the main thread. I do not have insight into the particular problem, but this is the first gut feel about this code, and it may be that iOS4 handles the (non-supported) mechanism of running UI functions in background threads somehow differently.
If you're using iOS 4.0, you should really consider reading up on blocks and GCD. Using those technologies, you can simply replace your method with:
- (void)decodeImageName:(NSString *)name
{
NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init];
UIImage *newImage = [UIImage imageNamed:name];
dispatch_async(dispatch_get_main_queue(), ^{
[myImageView setImage:newImage];
}
[pool release];
}
Let's quote:
#property(nonatomic, readonly) CGImageRef CGImage
Discussion
If the image data has been purged because of memory constraints, invoking this method forces that data to be loaded back into memory. Reloading the image data may incur a performance penalty.
So you might be able to just call image.CGImage. I don't think CGImages are lazy.
If that doesn't work, you can force a render with something like
// Possibly only safe in the main thread...
UIGraphicsBeginImageContext((CGSize){1,1});
[image drawInRect:(CGRect){1,1}];
UIGraphicsEndImageContext();
Some people warn about thread-safety. The docs say UIGraphics{Push,Pop,GetCurrent}Context() are main-thread-only but don't mention anything about UIGraphicsBeginImageContext(). If you're worried, use CGBitmapContextCreate and CGContextDrawImage.