Mailman 3 March 2022 - Typing-sig

Opened and changed typing issues week 2022-W10
by Typing Bot 13 Mar '22

13 Mar '22

The following is an overview of all issues and pull requests in the typing repository on GitHub with the label 'topic: feature' that were opened or updated last week, excluding closed issues. --------------------------------------------------- The following issues and pull requests were updated last week: #1096 Introduce `typing.STRICTER_STUBS` opened by @not-my-profile https://github.com/python/typing/issues/1096 --------------------------------------------------- All issues and pull requests with the label 'topic: feature' can be viewed under the following URL: https://github.com/python/typing/issues?q=label%3A%22topic%3A+feature%22

1 0

Decorator for creating a callable type
by Serhiy Storchaka 13 Mar '22

13 Mar '22

typing.Callable allows you to create annotations for limited set of functions. In particularly it does not support optional or keyword arguments. In general case you need to create a protocol with the __call__ method with a specific signature. I tried to implement the Mark Shannon's idea about a decorator which converts a function into corresponding type, but failed with this. Seems my skills are not enough for this. Use case: @callable_type def Func(x: int) -> str: pass def func(x: int) -> str: return chr(x) func2: Func = func I tried: def callable_type(func: Callable[P, T]): class result(Protocol): def __call__(self, /, *args: P.args, **kwargs: P.kwargs) -> T: pass return result but it does not work, How can this be implemented? It would be very useful feature which we would use on work a lot.

5 4

Opened and changed typing issues week 2022-W09
by Typing Bot 06 Mar '22

06 Mar '22

The following is an overview of all issues and pull requests in the typing repository on GitHub with the label 'topic: feature' that were opened or updated last week, excluding closed issues. --------------------------------------------------- The following issues and pull requests were opened last week: #1096 Introduce `typing.STRICTER_STUBS` opened by @not-my-profile https://github.com/python/typing/issues/1096 --------------------------------------------------- The following issues and pull requests were updated last week: #956 Add a `StubOnlyBase` decorator for types that don't actually extend things, but do in the stubs (like `Generic`) opened by @KotlinIsland https://github.com/python/typing/issues/956 --------------------------------------------------- All issues and pull requests with the label 'topic: feature' can be viewed under the following URL: https://github.com/python/typing/issues?q=label%3A%22topic%3A+feature%22

1 0

Typing Summit Presentation Requests
by Shannon Zhu 04 Mar '22

04 Mar '22

Hi everyone! Alfonso and I are working on putting together the agenda for this year’s typing summit at PyCon (4/28). There are a few topics that attendees have requested, but that we don’t yet have a volunteer to organize a presentation and/or lead a discussion about. If any of these are topics you have context on and might be interested in speaking about at the summit, please reach out to us! Time slots usually range anywhere from 10 to 40 minutes, and the format can be mostly slides, mostly organized group discussions, or anything in between. * Formalizing the Python type system specification: identifying gaps, agreeing on fundamentals like the definition of Any, subtyping for primitives, and more. * Type inference: sharing experience using MonkeyType or other similar tooling to increase type coverage * Brainstorming continued improvements to type annotation readability and usability * An overview of PEP 647 (Type Guards), and discussion on strict type guards that could support type narrowing in the negative case * Tensor types and future extensions for variadic typing * Type coverage in top libraries – current status and how we can make this better * Standardizing recursive type alias support * Type checker implementations / interesting problems (ie., Eric’s Pyright talk at the last meetup) Finally, if you haven’t yet filled out the interest form<https://docs.google.com/forms/d/1B3NfJEIw4S3XdyyZbfgHp608y3AXlWwoRY-Klai-as…> but are planning to come, please do so ASAP! 😊 Looking forward to seeing everyone at the summit, Shannon & Alfonso

3 3

Question about a possible improvement to the behaviour of unspecified Generic arguments.
by Joshua Butt 01 Mar '22

01 Mar '22

Hello, in debugging some old code of mine i encountered an issue where an attribute i would have expected to have had a type instead was Any, causing type-checking to ignore a possible bug in my code. Consider the following example code: --- from typing import TypeVar, Sized T = TypeVar('T', bound=Sized) class Foo(Generic[T]): def __init__(self, value: T) -> None: self.value: T = value def takes_foo(foo: Foo) -> None: len(list(reversed(foo.value))) --- Testing with mypy and pyright, by default both will type-check, however this code is unsound, foo.value may not be Reversible. This happens because as PEP 484 [1] states: --- Additionally, Any is a valid value for every type variable Consider the following: def count_truthy(elements: List[Any]) -> int: return sum(1 for elem in elements if elem) This is equivalent to omitting the generic notation. --- So the type of foo.value is Any and from that point on, everything type-checks. So I ask this: Is there any reason why type-checkers treating omitted generics as the TypeVar bound, if set, rather than Any could be problematic? Doing so would remove the chance of errors such as these from ocurring in a lot of situations. Interested to hear thoughts on this, thanks. [1] - https://www.python.org/dev/peps/pep-0484/#generics

3 2

Typing-sig March 2022