I am currently playing with the Python 3.5 interpreter and found a very interesting behavior:
>>> (1,2,3,"a",*("oi", "oi")*3)
(1, 2, 3, 'a', 'oi', 'oi', 'oi', 'oi', 'oi', 'oi')
>>> [1,2,3,"a",*range(10)]
[1, 2, 3, 'a', 0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> ('aw','aw',*range(10),*(x**2 for x in range(10)))
('aw', 'aw', 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 4, 9, 16, 25, 36, 49, 64, 81)
>>> {"trali":"vali", **dict(q=1,p=2)}
{'q': 1, 'p': 2, 'trali': 'vali'}
>>> {"a",1,11,*range(5)}
{0, 1, 2, 3, 4, 11, 'a'}
I have never seen this either in the documentation, in the examples, or in any source code, despite several years of my Python experience. And I found it very useful.
And that seems logical to me in terms of Python grammar. Function arguments and a tuple can be parsed with the same or similar states.
Is this documented behavior? Where is this documented?
What versions of Python have this functionality?