BUG: FloatingArray.equals should consider NaNs in same location as equal

[jorisvandenbossche]

From #43930 (comment). Currently we have

In [24]: arr1 = pd.arrays.FloatingArray(np.array([1.0, np.nan, 2.0]), np.array([False, False, True]))

In [25]: arr1
Out[25]: 
<FloatingArray>
[1.0, nan, <NA>]
Length: 3, dtype: Float64

In [26]: arr2 = pd.arrays.FloatingArray(np.array([1.0, np.nan, 2.0]), np.array([False, False, True]))

In [27]: arr1.equals(arr2)
Out[27]: False

For nullable arrays, NAs in the same location are considered equal. For plain float arrays, we consider NaNs as equal. Although for FloatingArray we currently distinguish both NA and NaN, I think we should consider both as equal if in the same location, to preserve the goal of this function (NaN equal to NaN, and NA equal to NA, to be clear, not mixed).