关于CABasicAnimation无插值值时的动画行为疑问（Apple官方文档相关）

Great question—this is one of those nuanced Core Animation behaviors that even seasoned iOS/macOS devs often overlook or misunderstand. Let’s break down exactly what happens, step by step, with practical context.

Core Logic at a Glance #

When you create a CABasicAnimation without setting fromValue, byValue, or toValue, Core Animation automatically infers the animation’s start and end values from two critical layer states:

• Start value: The real-time value of your animation’s keyPath on the layer’s presentation layer at the exact moment you add the animation. This is the value currently being displayed on screen (even if another animation is in progress).
• End value: The current value of the same keyPath on the layer’s model layer when you add the animation. This is the "true" data value stored in the layer, not what’s currently visible.

Why This Matters (And How It Works in Practice) #

The presentation layer and model layer often diverge during animations. For example:

• If you run an animation to move a layer from (100,100) to (300,300) over 1 second, halfway through the animation the presentation layer’s position is (200,200), but the model layer’s position is already (300,300) (since you set it before starting the animation).

If you add a new, nil-value CABasicAnimation mid-animation (say, to update the target position to (500,500)):

1. You first update the model layer’s position to (500,500).
2. When you add the new animation, Core Animation grabs the presentation layer’s current position (200,200) as the start.
3. The animation smoothly transitions from (200,200) to (500,500) using your specified duration and timing function.

This is incredibly useful for seamless, interactive animations—you don’t have to manually track or query the presentation layer’s current value to avoid jarring jumps.

Key Details to Remember #

• You must update the model layer first: If you forget to set the model layer’s property to your desired end value, the start and end values will be identical, and no animation will occur. The animation relies on the model layer to define where it should end.
• Timing and curves still apply: Your animation’s duration, timingFunction, repeatCount, etc., all work exactly as they would if you’d explicitly set fromValue and toValue.
• Handles all property types: This logic works for any animatable property (position, opacity, transform, etc.) as long as the model layer’s value matches the property’s type (just like when you set explicit interpolation values).

Example Code #

Here’s a simple Swift example demonstrating this behavior:

// 1. Update the model layer to our target value
myLayer.position = CGPoint(x: 400, y: 400)

// 2. Create a CABasicAnimation without any interpolation values
let smoothAnimation = CABasicAnimation(keyPath: "position")
smoothAnimation.duration = 0.5
smoothAnimation.timingFunction = CAMediaTimingFunction(name: .easeInEaseOut)

// 3. Add the animation—Core Animation handles the rest
myLayer.add(smoothAnimation, forKey: "positionTransition")

If the layer was previously at (100,100), this animation will smoothly move it from (100,100) (presentation layer’s start value) to (400,400) (model layer’s target value).

What If There’s No Active Animation? #

If the layer is static (no ongoing animations), the presentation layer’s value matches the model layer’s value—until you update the model layer and add the nil-value animation. At that point, the presentation layer’s old value becomes the animation start, and the new model layer value becomes the end.

内容的提问来源于stack exchange，提问作者Pawel Jurczyk