Alex Waygood

> Understood. That's a shame, since it makes `Sentinel` use very cumbersome when using type checkers I think it was probably the correct decision for the PEP, since the PEP...

> Sure, but at that point I don't understand what's the point of adding Signature. If we don't care about type checking support I think the classic `flag = object()`...

This still crashes on master, but the traceback is now different following #12762. Running mypy on the original repro now produces: ``` (venv) C:\Users\alexw\coding\mypy>mypy test.py Traceback (most recent call last):...

Yet another repro for this crash was identified by @DetachHead in #13377, which I just closed as a duplicate of this issue: ```python from dataclasses import dataclass, field from typing...

I'm raising the priority of this issue, since it has been requested many times and there have been at least three duplicate issues.

The `psycopg` primer errors are because of https://github.com/python/typeshed/pull/8446 — the maintainer has been warned, and said he was cool with it. For most of the other primer errors, I think...

> Why? I have no idea. Probably due to some typeshed changes. I think @AlexWaygood was working with `Match` recently. Yeah, I'm 99% sure this is somehow due to a...

The major change to `re` in this typeshed sync is https://github.com/python/typeshed/commit/ea7bbbdad7abf05a3bd3feb25eef93b53165e2d9. But I can't figure out why this is making `teststubtest` think that `re.Match` expects no type arguments. @hauntsaninja, do...

> The major change to `re` in this typeshed sync is [python/typeshed@ea7bbbd](https://github.com/python/typeshed/commit/ea7bbbdad7abf05a3bd3feb25eef93b53165e2d9). But I can't figure out why this is making `teststubtest` think that `re.Match` expects no type arguments. @hauntsaninja,...

> message.pyi:13: error: Cannot determine type of "Any" (diff) Uh, that looks like https://github.com/python/typeshed/issues/7019... Maybe syncing typeshed again might help??