I'm not sure how to proceed. I want to animate a slight graphics changes. I need this to happen in a loop every 2 seconds.
I have some drawing code in a class within a drawrect method. I pass in a style parameter with a custom method initWithFrame.
Within my animation animateWithDuration block, I thought I could just removeFromSuperview then addSubview different styled instances.
I know this sounds like the wrong approach.
However, I wasn't sure how else to do this.
Can anyone give me some options / point me at an example ?
To your original question about whether you should remove and re-add the view, typically you wouldn't do that (as that's a little inefficient). I usually have a public property in my custom view that (a) the view controller can set; and that (b) is used by the drawRect method. I then implement a custom setter that not only saves the ivar backing the property, but calls setNeedsDisplay, too (which will trigger drawRect to be called for you).
You haven't described what is changing in the view, but let me give you a random example. Let's say you have a drawRect that draws a circle of a given radius. Then my UIView subclass might have a property for this CGFloat, e.g.:
#property (nonatomic) CGFloat radius;
And then I might implement a custom setter which will not only update the ivar, like usual, but also triggers the re-calling of drawRect:
- (void) setRadius:(CGFloat)newRadius
{
_radius = newRadius; //note, I set the ivar backing the property; do not try self.radius = ..., because that will call this same method with infinite recursion
[self setNeedsDisplay];
}
Then, when my view controller wants to update the view, it can simply do the following:
customView.radius = 27.5;
Or, if I want to animate the changing of this every two seconds, fading to the new position, every two seconds, I'd wrap the changing of this property in a transitionWithView block:
[UIView transitionWithView:customView
duration:0.3
options:UIViewAnimationOptionTransitionCrossDissolve
animations:^{
customView.radius = 27.5;
}
completion:nil];
Now, clearly this is a random example where this property was a CGFloat, but you can do this with any custom property that your drawRect needs to reflect the new rendering of the screen.
Related
In my program I have started doing all initialization of objects in the init method without setting a frame and then in layoutSubviews I set the frames for these objects to make sure that they are properly set.
Firstly is this proper practice to initialize all objects in the init function without a set frame and then in layoutSubviews set each of their frames. The reason for my concern is that it is called quite often.
So I have a UIView subclass where I call these methods in the layoutSubviews
- (void)layoutSubviews
{
[super layoutSubviews];
[self.filterSwitcherView setFrame:self.viewFrame];
[self.drawingView setFrame:self.viewFrame];
[self.textView setFrame:self.textViewFrame];
[self.colorPicker setFrame:self.colorPickerFrame];
}
This currently works fine and all the objects are set correctly, but the problem is in my colorPicker class when the user touches the screen I adjust the frame of the colorPicker and by doing so this method gets called from the subview colorPicker and it readjusts a frame that it shouldn't since it has been modified in the subview. The subview causes the superviews layoutSubview to be called and this is not what I need.
My question is, is there a way to stop this behavior from happening or should I not use layoutSubviews to set frames because I was told this is a better way of making views programmatically?
Off the top of my head, there's two ways to fix this. You can either move this code to where the view is initialized, either in init, initWithFrame:, or initWithCoder:, depending on which you're using. It's good practice to make a separate method to initialize everything for your view, and call it from all the init methods to make sure it's always initialized correctly no matter how you instantiate the view.
Alternatively, if you want to keep your code in layoutSubviews, in your #interface add a boolean to flag that the frames were already set
#interface MyView : UIView
#property (nonatomic, assign) BOOL framesAreSet;
#end
Then when you set your frames, check if you already did
- (void)layoutSubviews
{
[super layoutSubviews];
if (!_framesAreSet)
{
[self.filterSwitcherView setFrame:self.viewFrame];
[self.drawingView setFrame:self.viewFrame];
[self.textView setFrame:self.textViewFrame];
[self.colorPicker setFrame:self.colorPickerFrame];
_framesAreSet = YES;
}
}
Your issue is likely that your colorPicker class is handling the touch methods to adjust its own frame. Instead, you should handle the touch methods in colorPicker's superview class, and have that superview class adjust colorPicker's frame in response to the touches.
Also, I would recommend doing all UI initialization in initWithFrame:, not init. The reason is because calling init on UIView ends up calling initWithFrame:.
Only one day left on the bounty!
Can you have a custom animatable property work with the UIView animateWithDuration block syntax?
I've been working on creating my own implicit animation for a custom animatable property on a custom CALayer.
I have set the backing layer of a custom UIView and can use a CATransaction to animate the custom property (in the custom example I am changing the color of a circle with a custom property circleColor).
I know that UIView turns off animatable properties by returning NSNull from the actionForLayer:forKey: selector method on UIView. Then when the property is wrapped in the UIView animateWithDuration block the UIView will return the animation.
So this will work with backgroundColor, opacity, etc. However, my custom property circleColor still returns NSNull.
I hope I provided enough context to the question. Thanks!
So this question has been open for a long-time and I'm concluding that it isn't possible. The code below shows how I would turn on the custom property to be detected within the block. This method is overridden in UIView.
- (id<CAAction>)actionForLayer:(CALayer *)layer forKey:(NSString *)event {
NSLog(#"%s - event: %#", __PRETTY_FUNCTION__, event);
id<CAAction> action = [super actionForLayer:layer forKey:event];
if ([event isEqualToString:#"circleColor"] && (nil == action || (id)[NSNull null] == action)) {
return [CABasicAnimation animationWithKeyPath:event];
}
return action;
}
The code allows you to animate within the UIView block however there is no way for it to detect that it is within a UIView block. So if you use this code then it will even animate outside of the UIView block which defeats the purpose. I'll monitor for other people's responses, but the only options it seems is to make your own custom UIView animation block system which will allow you to add custom animatable properties. I would imagine that Facebook's Pop framework would be a great alternative solution.
When using Interface Builder, a lot of the time I'd set the Class property to one of my custom classes. Say I have a UIView in the nib and I set the class to my MyView class.
That's all well and good, even when loading the nib programmatically.
But, is there a way to set the Class of a UI control when programmatically loading a nib? The reason being that I want to sometimes use the same nib created in IB, but have it associated with a different Class.
I don't think there is a proper way to achieve that programmatically.
You could create a naked UIView object in IB and give it a tag or assign it to an IBOutlet for its identification so taht you can access it programmatically in viewDidLoad.
In viewDidLoad you access it and fetch its frame and superview and probably background colour, alpha, hidden status etc, so that you can set all those values in IB. Then remove the UIView from its superview and nil the property (if any) or release the object respectively (depending on ARC or not). Then create the UI Element that you need, assign it to the property (if any) add it to the superview of the former UIView, set its frame and properties accordingly and go from there.
(There are ways in Obj-C to change an object's class after its instanciation, but I would not recommend doing that in this case, especially when your subclasses come with additional properties and ivars.)
Normally you create a UIView object on your UIViewController's class XIB file using Interface Builder and then you use the Custom Class tool in the Identity Inspector to associate the UIView with a class (The class is a UIView class containing the code for drawing on the UIView object.). Then you have to use #property and #systhesize in your UIViewController class to hook the UIView (using the connections inspector) to your class. This method is OK but in certain circumstances it has limitations.
You can get around all of this pragmatically.
Create the UIView class that will be used to draw on a UIView object. In this class you create the context reference (CGContextRef) to give the drawing tools a context (where to draw) for such things as strings, lines, circles, etc. i.e.
CGContextRef gg = UIGraphicsGetCurrentContext();
In the UIViewController class, in the .h file you need to import a reference to your UIView class. Let's call it: DrawOnView
#import "DrawOnView.h"
Then in the brackets encompassing the #interface place this line:
UIView * draw; // draw can be changed to any name that suits your needs
Then in the .m part of the class inside the viewDidLoad method you need to insert this code:
// Make the size and location that suits your needs
// You can change this on the go in your code as needed, such as if the
// device orientation is changed.
draw = [DrawOnView alloc] initWithFrame:CGRectCreate(50, 50, 100, 200)];
// You can change the background color of the view, if you like:
[draw setBackGroundColor:[UIColor greenColor]];
// Now add the view to your primary view
[self.view addSubview:draw];
Now, in other parts of our program, you can call methods you have declared in the DrawOnView class and refresh (which calls the drawRect method, the primary entry point in your UIView class) by using this reference:
[(DrawOnView*) draw setNeedsDisplay];
This is very important. Do not use:
[draw setNeedsDisplay]; // This will not work!
Let's say you have other methods defined in DrawOnView and want to call them.
Here's an example method (in the .h file):
-(BOOL) wasHotSpotHit: (CGPoint) p;
The actual method could look like this (in the .m file):
-(BOOL) washHotSpotHit: (CGPont) p
{
if(CGRectContainsPont(hotspot.frame, p))
{
return true;
}
return false;
}
Use code like this:
if([(DrawOnView*) draw testIfSpotHit:p])
{
// Do something for when user touches hot spot.
}
Try using "object_setClass(id object, Class cls)" method of Objective c runtime. GoodLuck:)
I had this question when/where to create and initialize views that are created programatically, so I hope some discussions here will shed more light on this topic for me.
This slide:
says: "not to initialize something based on the geometry of the view in viewDidLoad" and suggests viewDidAppear.
Imagine my view controller has view. I want to add 10 dynamic UIButtons to it.
Shall I put the code like below to the viewDidAppear?
-(void) viewDidAppear:(BOOL)animated
{
...
UIButton *button1 = [[UIButton alloc] initWithFrame: rect1];
[self.view addSubview: button1];
UIButton *button2 = [[UIButton alloc] initWithFrame: rect2];
[self.view addSubview: button2];
...
}
But this creates the buttons each time the view is shown. Is it what we want?
On the other hand if I put the code in viewDidLoad slide suggest not to initialize geometry of these views there.
Or shall we create buttons in viewDidLoad and set their frames in viewDidAppear?
What approach do you usually take?
But this creates the buttons each time the view is shown. It's true.
So the best thing you can do is to add a boolean (lets name it isLaunched). You set it to FALSE in the method -(void)viewDidLoad
Then add a if condition in your -(void)viewDidAppear where you perform creation of buttons (or other stuff) and set the boolean to true at the end.
You should have something like that :
-(void)viewDidLoad
{
//some settings
isLaunched = FALSE;
}
-(void)viewDidAppear:(BOOL)animated
{
if(!isLaunched)
{
//creating and adding buttons
isLaunched = TRUE;
}
}
zbMax (and now Amar) offered good solutions to implement the view creations in viewDidAppear: I will provide the rational for doing this (over viewDidLoad).
It is pretty simple actually. In viewDidLoad none of the views are actually setup yet, so any attempt to set/create frames or bounds will be extremely inconsistent. Struts and springs (or autolayout) will take effect after this method which will create additional changes to your views. viewDidAppear: is the correct place to do this because you can now rely on existing views and setting frames.
Reason for not playing with the geometry in viewDidLoad is because view is still in the memory and not on the window. Once the view is put on the window, then you can specify geometry. That happens when viewDidAppear is called for your controller.
As recommended, you should do all the initialisation in viewDidLoad as this is one time task and need not be repeated. Hold references to the added subviews and give them appropriate frame in viewDidAppear.
When you are dealing with custom UIView and its subviews, layoutSubviews is the method you need to override in the custom view in order to rearrange the geometry of its subviews.
Hope that helps!
I have a TestView inherited from UIView, and everything is drawn on the screen using its drawRect method.
But if I don't want drawRect to clear the view before drawing, I used
self.clearsContextBeforeDrawing = NO;
self.opaque = NO; // also added because the doc says the rect
// passed to drawRect will be filled with
// background color first if it is set to YES
in the initWithFrame method.
The drawRect method is invoked by using
[self.view setNeedsDisplay];
in the ViewController event handler. (for touchesMoved events)
But still, everything is cleared before anything is drawn? How to make it work?
I think this answers your question. From the answer:
You cannot prevent the contents from being erased by doing the following:
[self setClearsContextBeforeDrawing: NO];
This is merely a hint to the graphics engine that there is no point in having it pre-clear the view for you, since you will likely need to re-draw the whole area anyway. It may prevent your view from being automatically erased, but you cannot depend on it.