Fix how `IsDataclass` handles nested dataclasses so they can be compared
successfully. Previously something like the following would not work:
```python
from dataclasses import dataclass
from dirty_equals import IsDataclass

@DataClass
class Address:
    street: str
    zip_code: str

@DataClass
class Person:
    name: str
    address: Address

person = Person(
    name='Alice',
    address=Address(street='123 Main St', zip_code='12345'),
)

assert person == IsDataclass(
    name='Alice',
    address=IsDataclass(street='123 Main St', zip_code='12345')
)
```

This is a result of `IsDataclass` converting dataclasses to a dictionary
for comparison using `dataclasses.asdict`, which _recursively_ converts
dataclasses to plain dictionaries. Thus the inner `IsDataclass` fails,
because the other side of the comparison is now a dictionary, not a
dataclass.

These changes fix this by shallowly converting the dataclass to a
dictionary using [the approach recommended in the dataclasses
documentation](https://docs.python.org/3/library/dataclasses.html#dataclasses.asdict):
```python
{field.name: getattr(obj, field.name) for field in fields(obj)}
```

These changes also include a test for nested dataclass comparisons to
ensure this works as expected and continues to work in the future.