问题

以下是翻译好的部分：

以下是翻译好的部分：

以下代码：
```python
class Foo(int):
    def __new__(cls, x, *args, **kwargs):
        x = x if isinstance(x, int) else 42
        
        return super(Foo, cls).__new__(cls, x, *args, **kwargs)

在PyCharm中产生警告：“期望类型为'str | bytes | bytearray'，但实际为'int'”在最后一行的x上。

为什么会这样？

如果我评估super(Size, cls).__new__ == int.__new__，结果是True。这不也期望一个int吗？

如果我想在首次分配值或将某个值强制转换为此类型时添加行为，是否有更好的方法来创建int的子类？

这样的类的具体示例是class Size(int)，可以实例化为Size(1024)或Size('1 KiB')，结果相同。

<details>
<summary>英文:</summary>

The following code:

class Foo(int):
def new(cls, x, *args, **kwargs):
x = x if isinstance(x, int) else 42

    return super(Foo, cls).__new__(cls, x, *args, **kwargs)

Results in a warning (in PyCharm): "`Expected type 'str | bytes | bytearray', got 'int' instead`" on `x` in the last line.

Why is this?

If I evaluate `super(Size, cls).__new__ == int.__new__`, the result is `True`. Wouldn't that expect an `int` as well?

Is there a better way to create a subclass of `int`, if I want to add behaviour when a value is first assigned or some value is cast to this type?

A concrete example of such a class would be a `class Size(int)` that could be instantiated as `Size(1024)` or `Size('1 KiB')`, with the same result.

</details>


# 答案1
**得分**: 2

The overload of `int` with multiple args takes a string as the first param. It is roughly equivalent to

```python
int(x:int = 0) -> int:
int(x:str, base: int = 10) -> int:

You are providing __new__(cls, x, *args, **kwargs), so it is expecting that there should be a string as the first param.

So for this to not complain, drop the extra args:

class Foo(int):
    def __new__(cls, x):
        x = x if isinstance(x, int) else 42
        return super(Foo, cls).__new__(cls, x)


print(Foo(2))
print(Foo("2"))

2
42

int(x:int = 0) -> int:
int(x:str, base: int = 10) -> int:

class Foo(int):
    def __new__(cls, x):
        x = x if isinstance(x, int) else 42
        return super(Foo, cls).__new__(cls, x)


print(Foo(2))
print(Foo("2"))

2
42