注：使用的是Python 2.7。
一個簡單實現
class Foo(object):
__instance = None
def __init__(self):
pass
@classmethod
def getinstance(cls):
if(cls.__instance == None):
cls.__instance = Foo()
return cls.__instance
if __name__ == '__main__':
foo1 = Foo.getinstance()
foo2 = Foo.getinstance()
print id(foo1)
print id(foo2)
print id(Foo())
輸出的前兩個結果是相同的（id(foo1)與id(foo2)的值相同），第三個結果和前兩個不同。這里類方法getinstance()用于獲取單例，但是類本身也可以實例化，這樣的方式其實并不符合單例模式的要求。但是這樣做也有好處，代碼簡單，大家約定好這樣子調用就行了。但是最好在類的命名上也體現了出來這是一個單例類，例如Foo_singleton。換一個思路
先說一下init和new的區別：
class Foo(object):
__instance = None
def __init__(self):
print 'init'
if __name__ == '__main__':
foo = Foo()
運行結果是：
init
而下面的示例：
class Foo(object):
__instance = None
def __init__(self):
print 'init'
def __new__(cls, *args, **kwargs):
print 'new'
if __name__ == '__main__':
foo = Foo()
運行結果是：
new
new是一個類方法，會創建對象時調用。而init方法是在創建完對象后調用，對當前對象的實例做一些一些初始化，無返回值。如果重寫了new而在new里面沒有調用init或者沒有返回實例,那么init將不起作用。以下內容引用自http://docs.python.org/2/reference/datamodel.html#object.new
If __new__() returns an instance of cls, then the new instance's __init__() method will be invoked like __init__(self[, ...]), where self is the new instance and the remaining arguments are the same as were passed to __new__().
If __new__() does not return an instance of cls, then the new instance's __init__() method will not be invoked.
這樣做：
class Foo(object):
__instance = None
def __init__(self):
print 'init'
def __new__(cls, *args, **kwargs):
print 'new'
if cls.__instance == None:
cls.__instance = cls.__new__(cls, *args, **kwargs)
return cls.__instance
if __name__ == '__main__':
foo = Foo()
錯誤如下：
RuntimeError: maximum recursion depth exceeded in cmp
而這樣也有一樣的錯誤：
class Foo(object):
__instance = None
def __init__(self):
if self.__class__.__instance == None:
self.__class__.__instance = Foo()
print 'init'
if __name__ == '__main__':
foo = Foo()
該怎么做呢？下面參考了http://stackoverflow.com/questions/31875/is-there-a-simple-elegant-way-to-define-singletons-in-python/31887#31887：
class Foo(object):
__instance = None
def __new__(cls, *args, **kwargs):
print 'hhhhhhhhh'
if not cls.__instance:
cls.__instance = super(Foo, cls).__new__(cls, *args, **kwargs)
return cls.__instance
def hi(self):
print 'hi, world'
print 'hi, letian'
if __name__ == '__main__':
foo1 = Foo()
foo2 = Foo()
print id(foo1)
print id(foo2)
print isinstance(foo1, object)
print isinstance(foo1, Foo)
foo1.hi()
運行結果：
hhhhhhhhh
hhhhhhhhh
39578896
39578896
True
True
hi, world
hi, letian
那么，到底發生了什么，我們先回顧一下super：
>>> print super.__doc__
super(type) -> unbound super object
super(type, obj) -> bound super object; requires isinstance(obj, type)
super(type, type2) -> bound super object; requires issubclass(type2, type)
Typical use to call a cooperative superclass method:
class C(B):
def meth(self, arg):
super(C, self).meth(arg)
可以肯定上面的單例模式代碼中的這一行代碼：
cls.__instance = super(Foo, cls).__new__(cls, *args, **kwargs)
super(Foo, cls)是object，super(Foo, cls).new方法使用的是object的new方法。我們看一下object.new方法的作用：
>>> print object.__new__.__doc__
T.__new__(S, ...) -> a new object with type S, a subtype of T
如果是一個繼承鏈
class Fo(object):
def __new__(cls, *args, **kwargs):
print 'hi, i am Fo'
return super(Fo, cls).__new__(cls, *args, **kwargs)
class Foo(Fo):
__instance = None
def __new__(cls, *args, **kwargs):
if not cls.__instance:
print Foo is cls
print issubclass(cls, Fo)
print issubclass(cls, object)
cls.__instance = super(Foo, cls).__new__(cls, *args, **kwargs)
return cls.__instance
def hi(self):
print 'hi, world'
if __name__ == '__main__':
foo1 = Foo()
foo1.hi()
print isinstance(foo1, Foo)
print isinstance(foo1, Fo)
print isinstance(foo1, object)
運行結果如下：
True
True
True
hi, i am Fo
hi, world
True
True
True
如果如下定義Fo，也正常運行：
class Fo(object):
pass
但是，若這樣定義：
class Fo(object):
def __new__(cls, *args, **kwargs):
print 'hi, i am Fo'
運行時報錯如下：
AttributeError: 'NoneType' object has no attribute 'hi'
