Anchor Point in CALayer

Looking at the Touches example from Apple's documentation, there is this method:

// scale and rotation transforms are applied relative to the layer's anchor point
// this method moves a gesture recognizer's view's anchor point between the user's fingers
- (void)adjustAnchorPointForGestureRecognizer:(UIGestureRecognizer *)gestureRecognizer {
    if (gestureRecognizer.state == UIGestureRecognizerStateBegan) {
        UI开发者_Go百科View *piece = gestureRecognizer.view;
        CGPoint locationInView = [gestureRecognizer locationInView:piece];
        CGPoint locationInSuperview = [gestureRecognizer locationInView:piece.superview];

        piece.layer.anchorPoint = CGPointMake(locationInView.x / piece.bounds.size.width, locationInView.y / piece.bounds.size.height);
        piece.center = locationInSuperview;
    }
}

First question, can someone explain the logic of setting the anchor point in the subview, and changing the center of the superview (like why this is done)?

Lastly, how does the math work for the anchorPoint statement? If you have a view that has a bounds of 500, 500, and say you touch at 100, 100 with one finger, 500, 500 with the other. In this box your normal anchor point is (250, 250). Now it's ???? (have no clue)

Thanks!

The center property of a view is a mere reflection of the position property of its backing layer. Surprisingly what this means is that the center need not be at the center of your view. Where position is situated within its bounds is based on the anchorPoint which takes in values anywhere between (0,0) and (1,1). Think of it as a normalized indicator of whether the position lies within its bounds. If you were to change either the anchorPoint or the position, the bounds will adjust itself rather than the position shifting w.r.t to its superlayer/superview. So to readjust position so that the frame of the view doesn't shift one can manipulate the center.

piece.layer.anchorPoint = CGPointMake(locationInView.x / piece.bounds.size.width, locationInView.y / piece.bounds.size.height);

Imagine the original thing being where O is the touch point,

+++++++++++  
+ O       +         +++++++++++
+    X    +  -->    + X       +
+         +         +         +
+++++++++++         +         +
                    +++++++++++

Now we want this X to be at the point where the user has touched. We do this because all scaling and rotations are done based on the position/anchorPoint. To adjust the frame back to its original position, we set the "center" of the view to the touch location.

piece.center = locationInSuperview;

So this reflects in the view readjusting its frame back,

                    +++++++++++  
+++++++++++         + X       +
+ X       +  -->    +         +
+         +         +         +
+         +         +++++++++++
+++++++++++

Now when the user rotates or scales, it will happen as if the axis were at the touch point rather than the true center of the view.

In your example, the location of view might end up being the average i.e. (300, 300) which means the anchorPoint would be (0.6, 0.6) and in response the frame will move up. To readjust we move the center to the touch location will move the frame back down.

First question, can someone explain the logic of setting the anchor point in the subview, and changing the center of the superview (like why this is done)?

This code isn't changing the center of the superview. It's changing the center of the gesture recognizer's view to be the location of the gesture (coordinates specified in the superview's frame). That statement is simply moving the view around in its superview while following the location of the gesture. Setting center can be thought of as a shorthand way of setting frame.

As for the anchor point, it affects how scale and rotation transforms are applied to the layer. For example, a layer will rotate using that anchor point as its axis of rotation. When scaling, all points are offset around the anchor point, which doesn't move itself.

Lastly, how does the math work for the anchorPoint statement? If you have a view that has a bounds of 500, 500, and say you touch at 100, 100 with one finger, 500, 500 with the other. In this box your normal anchor point is (250, 250). Now it's ???? (have no clue)

The key concept to note on the anchorPoint property is that the range of the values in the point is declared to be from [0, 1], no matter what that actual size of the layer is. So, if you have a view with bounds (500, 500) and you touch twice at (100, 100) and (500, 500), the location in the view of the gesture as a whole will be (300, 300), and the anchor point will be (300/500, 300/500) = (0.6, 0.6).