人言落日是天涯，望极天涯不见家 wrote: > Please see the followed example: > class A: > def __init__(self): > pass > > class X: > def __init__(self): > n = 200 > if True: > j = 200 > m = j > k = A() > print m, j > > a = X() > # ?? what about the m, n and j? is it still alive? > del a > > -------------------------- > In C/C++, the life time of m,n and j was the nearest block. but > obviously, python doen't have this syntax, but I would like to know > that whether the life time of m, n, j is base on function range or > the object range. > > We can not access the m, n, and j from the outside of class X. Now I'm > writing a program base on the wxpython. In the __init__ function of > wx.Panel, I use normal varable(just like the m,n and j) created some > widgets. It could be show in the window. Does it indicated the life > time of varable m,n,j is base on the object range?
Python has no variables. It has objects, which can be bound to names. Each binding to a name will increase a reference counter. Each unbinding will decrease it. so a = SomeObject() b = a del a will result in the SomeObject-instance still be alive. But when you add del b it will be garbage collected. Now in your example A() bound to k will not survive the exit of the method, as that means that k goes out of scope, and the object is bound to - the A-instance - gets its reference-counter decreased, resulting in it being freed. The wxwidgets example though is a different thing. If the panel stores a reference to the object, e.g. via a list (being part of a list or dict also increases the reference count), it will be kept around. Diez -- http://mail.python.org/mailman/listinfo/python-list
