Get pointer targets in environments that support layout

Problem description

An increasing number of our users applies this library in test environments which support layout.
In those environments given the clientX/clientY coordinates of a pointer device, we could detect the target of pointer interactions automatically and improve the confidence our users get from their tests.

Suggested solution

If a user supplies x/y coordinates for pointer interactions, determine related DOM node(s) and position automatically.

• Retrieve the element target per Document.elementFromPoint().
• Retrieve the selection position per experimental Document.caretPositionFromPoint() or Document.caretPositionFromPoint().

If a target and or node/offset is given and contradicts the determined element/position, throw an error.

Todo

It would be great if we could support not declaring a target in the API call.
So far we reuse the previous pointer position for subsequent pointer actions. This begs the question how we would determine if we keep the coordinates and/or the target as a reference point.

Start by tracing the pointer interaction API described in the issue and review Document.elementFromPoint(), caretPositionFromPoint(), and caretRangeFromPoint(). Done means x/y coordinates determine the element and selection position, contradictory target or node/offset values error, and subsequent pointer actions have defined coordinate and target behavior.