Modifies followObject in-place. Invoked after position is resolved each frame.
Called once at drag start and again whenever the drag context resets (e.g. multi-touch → single-pointer transition). Capture any object snapshot (position, rotation, scale) you need to hold fixed during the drag. Constraints that need no dynamic initialization may omit this method.
Used by SnapToSurfacesDragStrategy for two-pointer (multi-touch) drags. Casts a downward ray each frame from the follow object's world position and adjusts it so the dragged object's bounding-box contact face rests on the detected surface.
Only injected into the constraint pipeline for multi-touch; the single-pointer path continues to use SnapToSurfacesDragStrategy.update (pointer ray + drag plane intersection) unchanged.