Improve urlunsplit and urlunparse

[AlexWaygood]

• The first overload for urlunsplit and urlunparse is currently pointless. If you provide a tuple of a different length to these functions, type checkers simply fall back to the second overload, and don't emit an error: https://mypy-play.net/?mypy=latest&python=3.11&gist=966b9d881ce16ee67e220bff46d5f21c.
• Mypy currently infers str for urlunsplit((None, None, None, None, None)), but the returned value is b"": https://mypy-play.net/?mypy=latest&python=3.11&gist=dffd33f24d5e90b18cee1757d48149c6. Replace the first overload with an overload for Iterable[None].
• Use Iterable instead of Sequence, since these functions accept any iterable that can be unpacked; it doesn't have to be a sequence.

[github-actions]

Diff from mypy_primer, showing the effect of this PR on open source code:

werkzeug (https://github.com/pallets/werkzeug)
+ src/werkzeug/routing/map.py:763: error: Unused "type: ignore[type-var]" comment
+ src/werkzeug/routing/map.py:764: error: Argument 1 to "urlunsplit" has incompatible type "Tuple[str, str, str, Union[Mapping[str, Any], str, None], None]"; expected "Iterable[None]"  [arg-type]

[hauntsaninja]

Thanks, werkzeug error looks clearer (also I think the code might be genuinely buggy if someone passes empty dict)

Note, since we're caring more about the IDE setting now, this annotation will be a little less helpful, but I'm fine with it.