Python与抽象方法的不同行为 [英] Python different behaviour with abstractmethod
问题描述
我有两个类继承自同一个父P
:
I have two classes inheriting from the same parent P
:
from abc import ABCMeta, abstractmethod
class P(object):
__metaclass__ = ABCMeta
@abstractmethod
def foo(self):
pass
class C(P):
pass
class D(tuple, P):
pass
唯一的区别是D
继承自tuple
和P
,而C
继承自P
仅.
The only difference is that D
inherited from tuple
and P
while C
inherits from P
only.
现在这是行为:c = C()
出现错误,正如预期的那样:
Now this is the behavior: c = C()
got error, as expected:
TypeError: Can't instantiate abstract class C with abstract methods foo
但是 d = D()
可以正常工作!
but d = D()
works without error!
我什至可以调用 d.foo()
.我该如何解释这种行为?
I can even call d.foo()
. How can I explain this behaviour?
推荐答案
抽象方法在 object.__new__
方法中进行测试;当您从具有自己的 __new__
方法的 tuple
继承时,不会调用 object.__new__
并且不会对抽象方法进行测试.
Abstract methods are tested for in the object.__new__
method; when you inherit from tuple
, which has its own __new__
method, object.__new__
is not called and the test for abstract methods is not made.
换句话说,将抽象方法与任何的内置不可变类型混合会导致这个问题.
In other words, mixing abstract methods with any of the built-in immutable types will cause this problem.
唯一有效的解决方案是在 __new__
中进行自己的 测试,然后仅当您将抽象类放在 之前元组
在子类中混合两个基时.
The only solution that works is to do your own test in __new__
and then only if you put your abstract class before tuple
when mixing in the two bases in a subclass.
class P(object):
__metaclass__ = ABCMeta
def __new__(cls, *args, **kwargs):
super_new = super(P, cls).__new__
if super_new.__self__ is not object:
# immutable mix-in used, test for abstract methods
if getattr(cls, '__abstractmethods__'):
raise TypeError(
"Can't instantiate abstract class %s "
"with abstract methods %s" % (
cls.__name__,
', '.join(sorted(cls.__abstractmethods__))))
return super_new(cls, *args, **kwargs)
@abstractmethod
def foo(self):
pass
class D(P, tuple):
pass
这篇关于Python与抽象方法的不同行为的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!