python -- why defined '__new__' and '__init__' all in a class

粘贴解答至此：

You can define either or both of ​​__new__​​​ and ​​__init__​​.

​​__new__​​​ must return an object -- which can be a new one (typically that task is delegated to ​​type.__new__​​), an existing one (to implement singletons, "recycle" instances from a pool, and so on), or even one that's not an instance of the class. If ​​__new__​​​ returns an instance of the class (new or existing), ​​__init__​​​ then gets called on it; if ​​__new__​​ returns an object that's not an instance of the class, then ​​__init__​​ is not called.

​​__init__​​​ is passed a class instance as its first item (in the same state ​​__new__​​ returned it, i.e., typically "empty") and must alter it as needed to make it ready for use (most often by adding attributes).

In general it's best to use ​​__init__​​​ for all it can do -- and ​​__new__​​​, if something is left that ​​__init__​​ can't do, for that "extra something".

So you'll typically define both if there's something useful you can do in ​​__init__​​, but not everything you want to happen when the class gets instantiated.

For example, consider a class that subclasses ​​int​​​ but also has a ​​foo​​​ slot -- and you want it to be instantiated with an initializer for the ​​int​​​ and one for the ​​.foo​​​. As ​​int​​​ is immutable, that part has to happen in ​​__new__​​, so pedantically one could code:

>>>class x(int):...def __new__(cls, i, foo):...     self = int.__new__(cls, i)...return self
...def __init__(self, i, foo):...     self.foo = foo
...   __slots__ ='foo',...>>> a = x(23,'bah')>>>print a
23>>>print a.foo
bah
>>>

In practice, for a case this simple, nobody would mind if you lost the ​​__init__​​​ and just moved the ​​self.foo = foo​​​ to ​​__new__​​​. But if initialization is rich and complex enough to be best placed in ​​__init__​​, this idea is worth keeping in mind.