Misc questions, suggestions, and error corrections for dataclasses

[rhettinger]

These are for Eric's consideration:

• Fix typo leading to a SyntaxError in the make_dataclass docstring
• Let make_dataclass pass through keyword arguments (ordered, hash, frozen, etc)
• Replace "len(s)==0" with the faster and more idiomatic (PEP 8ish) "not s"
• Ask whether asdict() should return an OrderedDict instead of a regular dict
• Even though the class implementation details have "hash=None" as a way of turning off hashing, the dataclass API should use regular booleans and have the default be "hash=False" (as equivalent to "hash=None"). It is a bit of an API landmine to have hashing succeed in this example:

>>> from dataclasses import dataclass
>>> @dataclass(hash=False)
... class C:
...     x: int
...
>>> hash(C(999))
-9223372036585142703

[ericvsmith]

Thanks, @rhettinger. @serhiy-storchaka had many of these suggestions over at #4704, and I've been working on a patch for his issues. I'll answer your questions when I have a few minutes to devote to it, but it will probably have to wait until the weekend.

[ericvsmith]

Of the original issues, I think most have been addressed in subsequent commits (many not by me).

The remaining issues are:

• use "if fields" instead of "if len(fields) == 0"
• an issue with hashing.

I'm still investigating all of the hashing implications, and how it interacts with bpo-32513.

For the "if fields" issue, I'll just go ahead and commit that when I'm making other changes to dataclasses. So, I think this PR can be closed. I'll close it once I'm done with the hashing questions.