fix: Indent of generated code for non-required lists. Thanks @sfowl!

[sfowl]

The _transform macro is called at three different places in this file, but one location was not similarly indented as the other two. This caused errors in the ruff post_hook like:

error: Failed to parse foo/models/foo_request.py:125:9: Expected an indented block after `if` statement

The generated code without the indent fix looks like:

        if not isinstance(self.foo, Unset):
        _temp_foo = []  # <-- incorrect indent
            for foo_item_data in self.foo:
                foo_item: Union[None, int]
                foo_item = foo_item_data
                _temp_foo.append(foo_item)
            foo = (None, json.dumps(_temp_foo).encode(), 'application/json')

[sfowl]

I believe this bug was introduced by #926

[dbanty]

Can you provide a minimal OpenAPI schema which triggers this issue? That way we can make sure we're exercising that code path in tests to prevent this regression.

[sfowl]

I've modified the baseline spec in the end to end tests to cover this case, and regenerated the snapshots with pdm regen.

[dbanty]

Great, thanks so much!