It is actually possible to pass arrays and lists of booleans to Series objects that are not the same length as the DataFrame you are doing the indexing on. Let's look at an example of this by selecting the first and third rows, and the first and fourth columns:
>>> movie.loc[[True, False, True], [True, False, False, True]]
Both of the boolean lists are not the same length as the axis they are indexing. The rest of the rows and columns not explicitly given a boolean value in the lists are dropped.