CALayer setNeedsDisplayInRect - what is the point of InRect? Or faster drawing in drawInContext

I have a CALayer (imageLayer), and a subclassed CALayer that acts as a mask on the CALayer (maskLayer).

I am painting with a finger onto the maskLayer so that I can temporarily erase out part of the imageLayer.

I am accomplishing this by keeping one Bitmap Context (maskContext), and adding a stroke to it in touchesMoved. So my maskContext just holds all the strokes I have painted. In touchesMoved, I do [maskLayer setNeedsDisplay];

in my subclassed maskLayer, I have the drawInContext method:

-(void)drawInContext:(CGContextRef)ctx
{
    CGImageRef image = CGBitmapContextCreateImage(maskContext);
    CGContextDrawImage(ctx, CGRectMake(0,0,self.frame.size.width, self.frame.size.height), image);
}

This all works out really well, but is very slow as it's drawing the whole view for each touchesMoved event.

I thought of using [maskLayer setNeedsDisplayInRect:rect] but I wondered what the point is, as drawInContext does not look at the rect. I could pass the rect to drawInContext, but I could do that with just setNeedsDisplay.

Is there something that setNeedsDisplayInRect does that setNeedsDisplay doesn't for a CALayer?

(I am aware that if I were using drawRect, that would accept the defined rect, but a CALayer's drawInContext doesn't)

Also, is there any better way of accomplishing this masking by painting with a finger? The main problem is that the CGContext is not a canvas and does not remember previous strokes, so they have to be开发者_StackOverflow redrawn all the time.

---

If you mark a layer to display by calling setNeedsDisplayInRect: then you can use CGContextGetClipBoundingBox to get the area marked to display in your drawInContext:.