I want to draw text onto my subclass on UIView so that the text is cut out of the shape and the background behind the view shows through, just like in the OSX Mavericks logo
Disclaimer: I'm writing this without testing, so forgive me if I'm wrong here.
You should achieve what you need by these two steps:
Create a CATextLayer
with the size of your view, set the backgroundColor
to be fully transparent and foregroundColor
to be opaque (by [UIColor colorWithWhite:0 alpha:1]
and [UIColor colorWithWhite:0 alpha:0]
. Then set the string
property to the string you want to render, font
and fontSize
etc.
Set your view's layer's mask to this layer: myView.layer.mask = textLayer
. You'll have to import QuartzCore
to access the CALayer
of your view.
Note that it's possible that I switched between the opaque and transparent color in the first step.
Edit: Indeed, Noah was right. To overcome this, I used CoreGraphics with the kCGBlendModeDestinationOut
blend mode.
First, a sample view that shows that it indeed works:
@implementation TestView
- (id)initWithFrame:(CGRect)frame {
if (self = [super initWithFrame:frame]) {
self.backgroundColor = [UIColor clearColor];
}
return self;
}
- (void)drawRect:(CGRect)rect {
[[UIColor redColor] setFill];
UIBezierPath *path = [UIBezierPath bezierPathWithRoundedRect:self.bounds cornerRadius:10];
[path fill];
CGContextRef context = UIGraphicsGetCurrentContext();
CGContextSaveGState(context); {
CGContextSetBlendMode(context, kCGBlendModeDestinationOut);
[@"Hello!" drawAtPoint:CGPointZero withFont:[UIFont systemFontOfSize:24]];
} CGContextRestoreGState(context);
}
@end
After adding this to your view controller, you'll see the view behind TestView
where Hello!
is drawn.
Why does this work:
The blend mode is defined as R = D*(1 - Sa)
, meaning we need opposite alpha values than in the mask
layer I suggested earlier. Therefore, all you need to do is to draw with an opaque color and this will be subtracted from the stuff you've drawn on the view beforehand.
If all you want is a white view with some stuff (text images, etc) cut out, then you can just do
yourView.layer.compositingFilter = "screenBlendMode"
This will leave the white parts white and the black parts will be see-through.
StatusReport's accepted answer is beautifully written and because I have yet to find a Swift answer to this question, I thought I'd use his answer as the template for the Swift version. I added a little extensibility to the view by allowing the input of the string as a parameter of the view to remind people that that his can be done with all of the view's properties (corner radius, color, etc.) to make the view completely extensible.
The frame is zeroed out in the initializer because you're most likely going to apply constraints. If you aren't using constraints, omit this.
Swift 5
class CustomView: UIView {
var title: String
init(title: String) {
self.title = title
super.init(frame: CGRect.zero)
config()
}
required init?(coder aDecoder: NSCoder) {
fatalError("init(coder:) has not been implemented")
}
private func config() {
backgroundColor = UIColor.clear
}
override func draw(_ rect: CGRect) { // do not call super
UIColor.red.setFill()
let path = UIBezierPath(roundedRect: bounds, cornerRadius: 10)
path.fill()
weak var context = UIGraphicsGetCurrentContext() // Apple wants this to be weak
context?.saveGState()
context?.setBlendMode(CGBlendMode.destinationOut)
title.draw(at: CGPoint.zero, withAttributes: [NSAttributedString.Key.font : UIFont.systemFont(ofSize: 24)])
context?.restoreGState()
}
}
let customView = CustomView(title: "great success")
I actually figured out how to do it on my own surprisingly but @StatusReport's answer is completely valid and works as it stands now.
Here's how I did it:
-(void)drawRect:(CGRect)rect{
CGContextRef context = UIGraphicsGetCurrentContext();
[[UIColor darkGrayColor]setFill]; //this becomes the color of the alpha mask
CGContextFillRect(context, rect);
CGContextSaveState(context);
[[UIColor whiteColor]setFill];
//Have to flip the context since core graphics is done in cartesian coordinates
CGContextTranslateCTM(context, 0, rect.size.height);
CGContextScaleCTM(context, 1.0, -1.0);
[textToDraw drawInRect:rect withFont:[UIFont fontWithName:@"HelveticaNeue-Thin" size:40];
CGContextRestoreGState(context);
CGImageRef alphaMask = CGBitmapContextCreateImage(context);
[[UIColor whiteColor]setFill];
CGContextFillRect(context, rect);
CGContextSaveGState(context);
CGContentClipToMask(context, rect, alphaMask);
[backgroundImage drawInRect:rect];
CGContextRestoreGState(context);
CGImageRelease(alphaMask);
}
- (void)setTextToDraw:(NSString*)text{
if(text != textToDraw){
textToDraw = text;
[self setNeedsDisplay];
}
}
I have an @synthesize
declaration for textToDraw
so I could override the setter and call [self setNeedsDisplay]
. Not sure if that's totally necessary or not.
I'm sure this has some typos but I can assure you, the spell checked version runs just fine.