I created an empty Cocoa app on Xcode for OS X, and added:
- (void)applicationDidFinishLaunching:(NSNotification *)aNotification {
self.view = [[NSView alloc] initWithFrame:NSMakeRect(100, 100, 200, 200)];
self.view.wantsLayer = YES;
self.view.layer = [CALayer layer];
self.view.layer.backgroundColor = [[NSColor yellowColor] CGColor];
self.view.layer.anchorPoint = CGPointMake(0.5, 0.5);
self.view.layer.transform = CATransform3DMakeRotation(30 * M_PI / 180, 1, 1, 1);
[self.window.contentView addSubview:self.view];
}
But the rotated layer's background is clipped by the view's bounding area:
I thought since some version of OS X and iOS, the view won't clip the content of its subviews and will show everything inside and outside? On iOS, I do see that behavior, but I wonder why it shows up like that and how to make everything show? (I am already using the most current Xcode 4.4.1 on Mountain Lion).
(note: if you try the code above, you will need to link to Quartz Core, and possibly import the quartz core header, although I wonder why I didn't import the header and it still compiles perfectly)
It turns out that if the line:
((NSView *)self.window.contentView).wantsLayer = YES;
is added to the very beginning, then it works as expected:
- (void)applicationDidFinishLaunching:(NSNotification *)aNotification
{
((NSView *)self.window.contentView).wantsLayer = YES;
self.view = [[NSView alloc] initWithFrame:NSMakeRect(200, 200, 200, 200)];
self.view.wantsLayer = YES;
self.view.layer.backgroundColor = [[NSColor yellowColor] CGColor];
[self.window.contentView addSubview:self.view];
self.view.layer.anchorPoint = CGPointMake(0.5, 0.5);
self.view.layer.transform = CATransform3DMakeRotation(30 * M_PI / 180, 0, 0, 1);
}
So it looks like if all the views are made to be layer backed, then it works the same as it does on iOS. (If there is a quick way to make all views layer backed automatically, that'd be good).
the anchorPoint line cannot be moved before addSubview line, or else it is incorrect, although I wonder why that would make any difference.
The line self.view.layer = [CALayer layer]; can be removed if window.contentView is layer backed. Both the contentView and self.view don't need to set the layer, and I wonder why too.
The transform line cannot be before the addSubview line, or else it won't rotate, and I wonder why too.
The third thing is that, I thought if I go to Interface Builder and make the contentView a class of ContentView (subclassing NSView), and in its init method, do a self.wantsLayer = YES;, then it would work too, but it does not.
But anyway, the code above works, and I will update the reasons above why when I find out more.
Related
What I have:
I have a UIView (named pView) which has as sublayer a CAGradientLayer. Practically is this:
ViewController -> View ->pView -> CAGradientLayer
This is the code that creates all this:
#implementation ViewController
- (void)viewDidLoad {
[super viewDidLoad];
UITestView * pView = nil;
UIColor * scolor = nil, *ecolor = nil;
self.view.backgroundColor = [UIColor yellowColor];
pView = [[UITestView alloc] initWithFrame: CGRectMake(self.view.frame.origin.x, 100.0, self.view.frame.size.width, 100.0)];
scolor = [UIColor colorWithRed:(14/255.0) green: (238/255.0) blue:(123/255.0) alpha:1];
ecolor = [UIColor colorWithRed:(6/255.0) green: (216/255.0) blue:(69/255.0) alpha:1];
// creating the gradient layer
CAGradientLayer * layer = [[CAGradientLayer alloc] init];
layer.frame = self.bounds;
layer.colors = #[(id)scolor.CGColor,(id)ecolor.CGColor];
[pView.layer insertSublayer:layer atIndex:0];
// creating a tapGestureRecognizer
[pView addGestureRecognizer:[[UITapGestureRecognizer alloc] initWithTarget:self action:#selector(handleTapInViews:)]];
[self.view addSubview:pView];
touched = NO;
}
....
#end
What I'm trying to do
Once a tap gesture is detected over pView I want to increase the height of pView by 100.0 but animatedly. If pView was previously touched (that is, its height was already increased by 100.0) I want to decrease the height of pView by 100.0 (that is, returning it to its original size);
What I already know
I know that since I want to change the frame of pView I must change also the frame (or bounds) of the CAGradientLayer attached to pView. Since I want to animate these changes, I want these animations occurs at the same time and have the same duration.
I know that the frame (or bounds) of a layer is only animatable inside an animation block.
What I do:
This is the option I've test:
- (void) handleTapInViews: (nonnull UITapGestureRecognizer *) sender {
pView = sender.view;
if (touched) {
[UIView animateWithDuration:0.4 animations:^{
pView.frame = CGRectMake(pView.frame.origin.x, pView.frame.origin.y, pView.frame.size.width, pView.frame.size.height - 100.0);
pView.layer.sublayers[0].frame = pView.bounds;
[self.view setNeedsLayout];
} completion:^(BOOL finished) {
touched = NO;
}];
}
else {
[UIView animateWithDuration:0.4 animations:^{
pView.frame = CGRectMake(pView.frame.origin.x, pView.frame.origin.y, pView.frame.size.width, pView.frame.size.height + 100.0);
pView.layer.sublayers[0].frame = pView.bounds;
[self.view setNeedsLayout];
} completion:^(BOOL finished) {
touched = YES;
}];
}
}
This option actually works (that is, both the frame of pView and the frame of the layer change animatedly) but they are not synchronised; that is, the changes in the frame of the layer are slightly (but perceptible) more faster than the changes in the frame of pView. This effect is more evident when the height of pView is decreased.
I 've also read about CABasicAnimation and CAAnimationGroup, in order to animate both the bounds and position of the layer. In this case also the animation of the layer is bit faster than the animation of the view (it is perceptible and it is not a matter of seconds in case anyone ask about setting duration or whatever) and also in this case, after the animation the bound of the layer return to its original size which is not the desired effect. I already know this last matter can be fixed assigned the new values to the layer at the end of the animation but I certainty do not know where in my code put that.
Most of what i've read regarding this other option is from these links:
link1
link2
link3
link4
In any case, does anybody please knows how can I fix this?? thanks in advance.
Well I ended setting the backgroundColor of Pview as ecolor = [UIColor colorWithRed:(6/255.0) green: (216/255.0) blue:(69/255.0) alpha:1. This way, the effect I commented about the animation of the layer been faster than the animation of the view is not noticeable now. I think maybe this is not the proper answer but it suit me.
I am using this code proposed by Bartosz to add a mask to an UIImageView. It works fine.
#import <QuartzCore/QuartzCore.h>
CALayer *mask = [CALayer layer];
mask.contents = (id)[[UIImage imageNamed:#"mask.png"] CGImage];
mask.frame = CGRectMake(0, 0, 320.0, 100.0);
yourImageView.layer.mask = mask;
yourImageView.layer.masksToBounds = YES;
In addition, I want to animate the mask, e.g. sliding the mask to the right, so that at the end of the animation, the mask is not applied to the UIImageView any more.
In my specific case, the mask uses a fully transparent image, so the UIImageView is not visible at the initial state (which works fine), but is expected to be so at the end of the animation. However, the idea may be reused to any other use case were masks need to be animated.
The idea is to manipulate the x-origin portion of the frame of the mask. So, I came up with this code:
[UIView animateWithDuration: 0.2
delay: 0
options: UIViewAnimationCurveEaseInOut
animations:^{
CGRect maskFrame = yourImageView.layer.mask.frame;
maskFrame.origin.x = 320.0;
yourImageView.layer.mask.frame = maskFrame;
}
completion:^(BOOL finished){}];
Unfortunately, the mask is applied to the whole UIImageView at any time, it's not sliding to the right.
UPDATE 1:
This is the code I am actually using the set up the view and mask: It's a UITableViewCell.
APPCell.m (APPCell.h "extends" UITableViewCell)
#import "APPCell.h"
#import <QuartzCore/QuartzCore.h>
#interface APPCell()
#property (strong, nonatomic) UIImageView *menu;
#property (strong, nonatomic) CALayer *menuMask;
...
#end
#implementation APPCell
...
- (id)initWithStyle:(UITableViewCellStyle)style reuseIdentifier:(NSString *)reuseIdentifier
{
self.menu = [[UIImageView alloc] initWithFrame:CGRectMake(0, 0, 320.0, 88.0)];
[self.menu setBackgroundColor:[UIColor clearColor]];
[self.menu setImage:[UIImage imageNamed:#"cell_back"]];
[self addSubview:self.menu];
self.menuMask = [CALayer layer];
self.menuMask.contents = (id)[[UIImage imageNamed:#"cell_mask"] CGImage];
self.menuMask.frame = CGRectMake(0, 0, 320.0, 88.0);
self.menu.layer.mask = self.menuMask;
self.menu.layer.masksToBounds = YES;
}
...
Instead of animating with the help of UIKit, I am now using implicit animation of CoreAnimation to move the mask layer:
APPCell.m
...
- (void)swipeLeft
{
self.menuMask.position = CGPointMake(-320.0, 0.0);
}
...
I can confirm that swipeLeft is called. I expect the mask "to be gone" and to see the [UIImage imageNamed:#"cell_back"]], which I do when I uncomment self.menu.layer.mask = self.menuMask.
Solution:
Instead of setting the content on the CALayer, I set the background color to white. This is the code I am using:
self.menuSubMenuMask = [CALayer layer];
self.menuSubMenuMask.backgroundColor = [[UIColor whiteColor] CGColor];
self.menuSubMenuMask.frame = CGRectMake(320.0, 0.0, 320.0, 88.0);
self.tableCellSubMenu.layer.mask = self.menuSubMenuMask;
self.tableCellSubMenu.layer.masksToBounds = YES;
In order to show the UIImageView the CALayer is applied to, the CALayer must NOT be "above" the UIImageView.
Animation with UIKit of UIViews is much more limited than using Core Animation directly. In particular what you are trying to animate is not one of animatable properties of a UIView. In addition as clarified in the View Programming Guide for iOS:
Note: If your view hosts custom layer objects—that is, layer objects without an associated view—you must use Core Animation to animate any changes to them.
This is the case in your example. You have added a CALayer to your view and UIKit will not be able to animate the result for you. On the other hand you can use Core Animation directly to animate the motion of your mask layer. You should be able to do this easily using implicit animation as described in the Core Animation Programming Guide. Please note that from the list of CALayer Animatable Properties that frame is not animatable. Instead you should use position.
You can achieve something you want by using CATransition, although this might not be the solution you want:
1) At first, set mask for your layer just as you did
2) When you want to remove mask and reveal your image, use the following code:
CATransition* transition = [CATransition animation];
transition.type = kCATransitionPush;
transition.subtype = kCATransitionFromRight;
transition.duration = 1.0;
[mask addAnimation:transition forKey:kCATransition];
imageView.layer.mask.contents = [UIImage imageNamed:#"black.png"].CGImage;
The main trick here - we created transition animation for our mask layer, so this animation will be applied when you change any (i'm not sure about any) property of mask layer. Now we set mask's content to completely black image to remove masking at all - now we've got smooth pushing animation where our masked image is going to the left and unmasked image is getting into its place
The easiest way is to use CoreAnimation itself:
CGPoint fromPoint = mask.position;
CGPoint toPoint = CGPointMake(fromPoint.x*3.0, fromPoint.y);
mask.position = toPoint; // CoreAnimation animations do *not* persist
CABasicAnimation *animation = [CABasicAnimation animationWithKeyPath:#"position"];
animation.fromValue = [NSValue valueWithCGPoint:fromPoint];
animation.toValue = [NSValue valueWithCGPoint:toPoint];
animation.duration = 4.0f;
[mask addAnimation:animation forKey:#"position"];
So this is my first ever attempt at using a CALayer. Build is successful and no reported bugs, so I assume I must be doing something obviously wrong. But the layer does not display at all.
- (void)viewDidLoad
{
// Get Reliant Magenta in amazingly verbose manner
CGColorSpaceRef rgbaColorSpace = CGColorSpaceCreateDeviceRGB();
CGFloat reliantMagentaValues[4] = {(208/255),(27/255),(124/255),0.3f};
CGColorRef reliantMagenta = CGColorCreate(rgbaColorSpace, reliantMagentaValues);
CALayer *reliantCanvasLayer = [CALayer layer];
reliantCanvasLayer.backgroundColor = reliantMagenta;
reliantCanvasLayer.frame = CGRectMake(0, 0, 640, 960);
[super viewDidLoad];
[[[self view] layer] addSublayer:reliantCanvasLayer];
CGColorRelease(reliantMagenta);
}
Instead of a full page of magenta, I get back an empty view of grey. How am I messing up something this simple?
UPDATE
- (void)viewDidLoad
{
[super viewDidLoad];
// Get Reliant Magenta in amazingly verbose manner
CGColorSpaceRef rgbaColorSpace = CGColorSpaceCreateDeviceRGB();
CGFloat reliantMagentaValues[4] = {(208/255),(27/255),(124/255),0.3f};
CGColorRef reliantMagenta = CGColorCreate(rgbaColorSpace, reliantMagentaValues);
[[self view] layer].backgroundColor = reliantMagenta;
CGColorRelease(reliantMagenta);
}
same problem, but view is now black and not displaying elements added in the storyboard
One problem (possibly the only problem) is that you're creating your color with all zero components. When you say 208/255, the compiler performs the division using integers and drops the remainder, so 208/255 is 0. You need to divide in floating point: 208.0f / 255.0f.
It's also much easier to use a UIColor instead of setting up the CGColorSpace and the CGColor yourself. Try this:
- (void)viewDidLoad {
[super viewDidLoad];
UIColor *reliantMagenta = [UIColor colorWithRed:208.0f / 255.0f
green:27.0f / 255.0f blue:124.0f / 255.0f alpha:0.3f];
CALayer *magentaLayer = [CALayer layer];
magentaLayer.frame = CGRectMake(0, 0, 640, 960);
magentaLayer.backgroundColor = reliantMagenta.CGColor;
[self.view.layer addSublayer:magentaLayer];
}
You're adding a layer as a sublayer of itself. [self layer] returns the view's existing layer. If you want to create a separate layer, you have to do it manually. There's probably something in the view system that keeps circular references from iterating out of control.
As a point of reference, you should make it a habit of calling [super viewDidLoad] before you do anything else. And an easier way of creating a CGColor is to create a UIColor and get its CGColor property.
Since the template of an OS X app in Xcode seems to be similar to an empty app template, the following is used to add a view and a button (trying not to use Interface builder for now):
- (void)applicationDidFinishLaunching:(NSNotification *)aNotification
{
NSView *view = [[NSView alloc] initWithFrame:NSMakeRect(100, 100, 100, 100)];
view.layer.backgroundColor = [[NSColor yellowColor] CGColor];
[self.window.contentView addSubview:view];
NSRect frame = NSMakeRect(10, 40, 90, 40);
NSButton* pushButton = [[NSButton alloc] initWithFrame: frame];
pushButton.bezelStyle = NSRoundedBezelStyle;
[self.window.contentView addSubview:pushButton];
NSLog(#"subviews are %#", [self.window.contentView subviews]);
}
Similar code on iOS should have produced a yellow box and a button, but the code above only produce a button, but the view won't show. Is there something wrong with the code above, and how to make it show the view with a yellow background?
Use setWantsLayer: method of NSView class.
NSView *view = [[NSView alloc] initWithFrame:NSMakeRect(100, 100, 100, 100)];
[view setWantsLayer:YES];
view.layer.backgroundColor = [[NSColor yellowColor] CGColor];
[self.window.contentView addSubview:view];
NSRect frame = NSMakeRect(10, 40, 90, 40);
NSButton* pushButton = [[NSButton alloc] initWithFrame: frame];
pushButton.bezelStyle = NSRoundedBezelStyle;
[self.window.contentView addSubview:pushButton];
NSLog(#"subviews are %#", [self.window.contentView subviews]);
To expand on the suggestion by Kevin Ballard, the classic way to do this is to subclass NSView and override the -drawRect: method. NSRectFill is a very convenient function for filling a rectangle without having to create a bezier path:
- (void)drawRect:(NSRect)rect
{
[[NSColor yellowColor] set];
NSRectFill(rect);
}
NSViews in Cocoa are, by default, not layer-backed. I suspect that if you type
NSLog(#"%#", view.layer);
you will see that it is nil.
In iOS, all views have layers. But on OS X, views don't have layers. In addition, there's 2 "modes" of layer-backed views on OS X. There's what's called a "layer-backed views" and a "layer-hosting view". A layer-backed view uses a CoreAnimation layer to cache drawn data, but you are not allowed to interact with the layer in any way. A layer-hosting view uses a CALayer that you explicitly provide, and you may mess with that layer all you want. However, with a layer-hosting view you may not add any subviews, or use the built-in NSView drawing mechanism. A layer-hosting view must only be used as the root of a CoreAnimation layer hierarchy.
Given all this, you should probably avoid using CoreAnimation at all for your view.
It's possible that an NSBox will do what you want. You can certainly set a fill color there, turn off the border, and set the style to custom. I'm just not 100% certain it will draw as a simple filled rectangle of color. Alternatively you can define your own NSView subclass that draws a color in -drawRect:.
When I put shadow etc. with CALayer my app is lagging when I double tap on the home button to see tasks running. I don't have any other lag, just when I double tap.
I call this method 20 times to put 20 images :
- (UIView *)createImage:(CGFloat)posX posY:(CGFloat)posY imgName:(NSString *)imgName
{
UIView *myView = [[UIView alloc] init];
CALayer *sublayer = [CALayer layer];
sublayer.backgroundColor = [UIColor blueColor].CGColor;
sublayer.shadowOffset = CGSizeMake(0, 3);
sublayer.shadowRadius = 5.0;
sublayer.shadowColor = [UIColor blackColor].CGColor;
sublayer.shadowOpacity = 0.8;
sublayer.frame = CGRectMake(posX, posY, 65, 65);
sublayer.borderColor = [UIColor blackColor].CGColor;
sublayer.borderWidth = 2.0;
sublayer.cornerRadius = 10.0;
CALayer *imageLayer = [CALayer layer];
imageLayer.frame = sublayer.bounds;
imageLayer.cornerRadius = 10.0;
imageLayer.contents = (id) [UIImage imageNamed:imgName].CGImage;
imageLayer.masksToBounds = YES;
[sublayer addSublayer:imageLayer];
[myView.layer addSublayer:sublayer];
return myView;
}
I have commented all my code except this, so I'm sure the lag comes from here. Also I've checked with the Allocations tools and my app never exceeded 1Mo. When I'm just putting images without shadow etc. everything works fine.
Try setting a shadowPath on the layer as well. It will need to be a rounded rect since you've got rounded corners on your layer.
CALayer has to calculate where it is drawing, and where to put the shadow, if it doesn't have a shadow path. This has a big effect on animation performance.
Another way to improve performance with CALayers is to set the shouldRasterize property to YES. This stores the layer contents as a bitmap and prevents it having to re-render everything.