PointF.set（x，y）与Java中新的PointF（x，y）有何不同？

Question

I am trying to implement a scroll for my Canvas View in an Android app.

I initialize a global

private PointF backgroundPosition = new PointF(0, 0);

and then the behavior of this

backgroundPosition.set(lastBackgroundPosition.x + (canvasMovingStartingPoint.x - event.getX()), lastBackgroundPosition.y + (canvasMovingStartingPoint.y - event.getY()));

and this

backgroundPosition = new PointF(lastBackgroundPosition.x + (canvasMovingStartingPoint.x - event.getX()), lastBackgroundPosition.y + (canvasMovingStartingPoint.y - event.getY()));

is different for some reason (the scroll of the first version is much faster!!!)

WHY?!!! I have spent the whole day for it!

Here is the complete code:

    public boolean onTouchEvent(@NonNull MotionEvent event, GII.AppState appState) {
    switch (event.getAction()) {
        case MotionEvent.ACTION_DOWN:
            pressedHere = true;
            canvasMovingStartingPoint.set((int) event.getX(), (int) event.getY());
            lastBackgroundPosition = backgroundPosition;
            break;
        case MotionEvent.ACTION_MOVE:
            if (pressedHere &&
                    Math.sqrt((canvasMovingStartingPoint.x - event.getX()) * (canvasMovingStartingPoint.x - event.getX()) +
                            (canvasMovingStartingPoint.y - event.getY()) * (canvasMovingStartingPoint.y - event.getY())) > 10)
                moving = true;
            if (pressedHere && moving) {
                backgroundPosition.set(lastBackgroundPosition.x + (canvasMovingStartingPoint.x - event.getX()), lastBackgroundPosition.y + (canvasMovingStartingPoint.y - event.getY()));
                //backgroundPosition = new PointF(lastBackgroundPosition.x + (canvasMovingStartingPoint.x - event.getX()), lastBackgroundPosition.y + (canvasMovingStartingPoint.y - event.getY()));
                checkBackground();
            }
            break;
        case MotionEvent.ACTION_UP:
            if (pressedHere && !moving) {
                click(event.getX(), event.getY());
            }
            moving = false;
            break;
        default:
            return false;
    }
    return true;
}

Uncommenting that line makes everything perfect, the question is why?

Answer 1

In the bottom example you are creating a new instance of PointF

In the top one your just change the instances value.

Creating instances can be much more costly in memory then just reusing them so You are redundantly creating excess objects It is a low cost, but you should avoid creating unnecessary objects. Same as you should avoid unnecessary anything in your code.

As another note, from my understanding the x,y values in a PointF are final so they cannot be changed. So it appears the 2nd option is the way you have to go with since the set method appears to only work with empty instances. If it slows your program down to much you could always store the values in just two floats.