GenConverter's dict_factory argument has no effect

• cattrs version: 1.10.0
• Python version: 3.10.0
• Operating System: Win10

Description

I was experimenting with ways to modify the unstructured data, and noticed that GenConverter accepts the dict_factory argument, and also passes it to Converter, but effectively ignores it since the generated code hardcodes a dict.

For example

import attrs, cattr

@attrs.define
class Example:
    a: int
    b: int

class ExampleDict(dict):
    def __repr__(self):
        return f"ExampleDict({super().__repr__()})"

    def __setitem__(self, key, value):
        if abs(value) > 1<<52:
            value = str(value)
        super().__setitem__(key, value)


foo = Example(42, 1189998819991197253)

print(cattr.Converter(dict_factory=ExampleDict).unstructure(foo))
# ExampleDict({'a': 42, 'b': '1189998819991197253'})

print(cattr.GenConverter(dict_factory=ExampleDict).unstructure(foo))
# {'a': 42, 'b': 1189998819991197253}

GenConverter returns a dict (and therefore the unchanged values), while Converter returns an instance of ExampleDict, as expected.

Start by locating GenConverter and the generated unstructure path, then compare its dict_factory handling with Converter. Reproduce the Example/ExampleDict case from the issue and verify that generated unstructure output uses the configured factory and preserves its value transformation.