Closes #2505: Update resolve_scalar_dtype to handle uint and bigint

[stress-tess]

This PR (closes #2505) updates resolve_scalar_dtype to correctly handle bigint and uint scalars. It also updates places where this is handled differently to just call that function.

[Ethan-DeBandi99]

Looks good to me

[jaketrookman]

Everything looks good to me. Could you just explain why we need to switch from int to int64 here.

[stress-tess]

Everything looks good to me. Could you just explain why we need to switch from int to int64 here.

@jaketrookman yeah! so previously our message.py was using type(val).__name__ to determine the dtype of scalars and the dtype of elements within a list. Now that we've updated it to use resolve_scalar_dtype, it returns int64 instead of int this is nice because it matches the pdarray dtypes, but also it allows for uint64 and bigint.

So previously if you looked at the logs 2**63 - 1, 2**63 + 1, and 2**65 + 1 would all have dtype: int. But now they'll have int64, uint64, and bigint respectively