Talin

Question: who is going to be implementing the reference implementation for `typing.sealed` and the changes to `dataclass`?

Is there a decision on this? I am ambivalent, I can see the argument either way.

Also, you could have: ``` case {"a": 1, "b": 2}: # match ignoring extra values case Exact({"a": 1, "b": 2}): # don't allow extra values ```

Looks pretty easy, I could probably do it. What behavior do we want for `__match_args_required__`? len(fields) - len(defaults)?

I was wondering if we should write down all of the standard library classes that need to have match supported added. I don't think this needs to go in the...

I opened a PR for this. https://github.com/brandtbucher/cpython/pull/5 The build is failing for some reason, but "make test" succeeds locally. It's only a 5-line change (including tests).

On a somewhat related note, I wonder if we need additional explanations of `__match_args__` and `__match_args_required__` in the PEP. Right now these are defined in the PEP operationally - that...

Looking at the PEP, I'm not sure it is completely clear what the behavior is when `__match_args_required__` is not present. (I initially assumed that it defaulted to `len(__match_args__)` but apparently...

Part of the inconsistency here is that tuple and namedtuple accept constructor arguments in different ways: tuple accepts a single positional argument containing an iterable, whereas namedtuple accepts positional arguments...

Another, slightly more expensive but flexible option is to change `__match__` so that it returns a tuple of: ``` (target, match_args, match_args_required) ``` In other words, the custom match function...