Handles the case of
    float[] x = float[] (1.0, 2.0, 3.0),
            y = float[] (1.0, 2.0, 3.0, 4.0);
where a shallow copy of the type arrayness from the left-most float[]
was getting used twice.