Describe the Bug

The graphql schema generated by Generic subclasses is incorrect.

System Information

• Operating system: Ubuntu 20.04.3 LTS
• Strawberry version (if applicable): 0.171.1

Additional Context

Consider this code.

class MyGeneric(Generic[T]):
    ...


@strawberry.type
class Child(MyGeneric[int]):
    attr: int


@strawberry.type
class Parent:
    child: Child
    children: list[Child]
    maybe_child: Child | None
    annotated_child: Annotated[Child, 1]

@strawberry.type
class Query:
        parent = strawberry.field(resolver=get_parent)

The generated schema is

type Parent {
    annotatedChild: Child!
    child: IntChild!
    children: [Child!]!
    maybeChild: Child
}

That is a problem, because I write the same object everywhere in my python code, but my graphql schema doesn't use the same name. It duplicates my types for no good reason. It makes the frontend code harder to write as the classes change depending on if they are contained/annotated.

Note that Child itself will be a Child (and not an IntChild) when used as the root object.

I think the behaviour should be to have Child everywhere. In a more generic way:

class GenericChild(MyGeneric[T]):
    ...
 class Child(MyGeneric[int]):
 ...
 class Parent:
    generic_child : GenericChild[MyGeneric[int]]
   concrete_child: Child

should give:

type Parent {
   genericChild: IntGenericChild # This is a true generic which needs to be given details
   concreteChild: Child # This is not a generic anymore and doesn't need details
}

What do you say?