from __future__ import annotations

Question:

Python doc __future__

In python doc about __future__ there is a table below where it shows that
annotations "optional in" 3.7.0b1 and "mandatory in" 4.0
but I am still able to use annotations in 3.8.2 without importing annotations then what is the use of it.

>>> def add_int(a:int, b:int) -> int:
...     return a + b
>>> add_int.__annotations__
{'a': <class 'int'>, 'b': <class 'int'>, 'return': <class 'int'>}

I doubt I don’t clearly understand the meaning of "optional in" and "mandatory in" here

Answers:

Mandatory is an interesting word choice. I guess it means that it’s by default in the language. You don’t have to enable it with from __future__ import annotations

The annotations feature are referring to the PEP 563: Postponed evaluation of annotations. It’s an enhancement to the existing annotations feature which was initially introduced in python 3.0 and redefined as type hints in python 3.5, that’s why your code works under python 3.8.

Here’s what optional from __future__ import annotations changes in python 3.7+:

class A:
    def f(self) -> A: # NameError: name 'A' is not defined
        pass

but this works

from __future__ import annotations

class A:
    def f(self) -> A:
        pass

See this chapter in python 3.7 what’s new about postponed annotations:

Since this change breaks compatibility, the new behavior needs to be enabled on a per-module basis in Python 3.7 using a __future__ import:

from __future__ import annotations

It will become the default in Python 3.10*.

* it was announced to be default in 3.10 (when python3.7 was released), but it was now moved to a later release

Answered By: RafalS

Mandatory as in coming by default.
Optional as in needed to be "activated" from an from __future__ import annotations statement

Answered By: theX
Categories: questions Tags: , ,
Answers are sorted by their score. The answer accepted by the question owner as the best is marked with
at the top-right corner.