David Salvisberg

@AA-Turner Would you consider pulling in the backported version of Enum linked above for Python 2.5-2.7? At this point it seems better to force an additional dependency on people still...

@AA-Turner I personally disagree that it would only increase expressability. At least half of the point of static type checking is to find and report invalid arguments reliably. While technically...

If you're worried about maintenance burden, a reasonable compromise could be to only use `deprecated` in simple cases, such as renaming/removing a method/function/class. In these cases you would have to...

Note that as part of #79 we ~possibly~ temporarily dropped the required coverage percentage ~or omitted the test code from test coverage~, so this should once again be raised/~re-enabled~ once...

Here is an example tox.ini from pytest-codecov: https://github.com/seantis/pytest-codecov/blob/master/tox.ini Basically you are always measuring coverage for every test environment in a coverage file specific to that environment and all the report...

I recommend running `stubtest` which is included as part of `mypy` to automatically detect discrepancies between the stubs and the runtime package, that way changes can be done much more...

I think it might still be nice to have some support for stricter type checker configs on a per distribution basis on the CI side, but I agree that the...

This is already handled by the `__getattr__` method, so you technically don't need to explicitly add those attributes. As long as this magic method has been defined, any attributes that...

Also it seems to me like all of these should actually be `str | None`, since it's basically just using the result from `urllib.parse.parse_qsl` to initialize the params (but it...

@jvanasco Sure, you can could technically pass in an iterable of (key, value) tuples or a mapping as the body and set `encoding` to `None`, then it would bypass the...