According to http://bugs.python.org/msg176429 from http://bugs.python.org/issue12370:
class C(A, B, metaclass=meta):
def f(self):
return __class__To be compiled approximately like this:
def _outer_C(*__args__, **__kw__):
class _inner_C(*__args__, **__kw__):
def f(self):
return __class__
__class__ = _inner_C
return _inner_C
C = _outer_C(A, B, metaclass=meta)My question is:
class A:
def foo(a):
class X:
x = __class__
def fun(self):
print(__class__)
return X
print(A().foo().x)Why does this work? While the following doesn't?
def bar():
class Y:
def fun(a):
class X:
m = n
def foo(a):
print(a.m)
n = 1
return X
n = 0
return Y
if __name__ == '__main__':
y = bar()
x = y().fun()
x().foo()If the free variable is assigned in the outscope, it's turned to a cell var,
and because it's assigned after the class body is evaluated, NameError is raised.
If the __class__ closure trick is really what it said to be, it should also raise a NameError,
what's the trick use to make it work?