Leam Hall wrote:
> I do not understand your last sentence about reference cycle.
Currently you have
- create Career instance which stores character as an attribute
- make modifications to character
- forget about Career instance
My suggestion
- create Career instance which stores character as an attribute
- make modifications to character, among those
- enter career into the character.careers dict
Then the cycle goes
character --> careers dict --> career instance --> character
Such a cycle isn't too bad, I just thought I'd mention it for the gallery ;)
It means that reference count will never go to 0:
>>> class A:
... def __del__(self): print("A says bye!")
...
>>> class B:
... def __del__(self): print("B says bye!")
...
Case 1, a --> b, no cycle:
>>> a = A(); a.b = B()
>>> del a
A says bye!
B says bye!
Case 2, a --> b --> a, reference cycle:
>>> a = A(); a.b = B(); a.b.a = a
>>> del a
At that point we can neither access a nor b, but as a holds a reference to b
and b holds a reference to a the reference count of both stays above 0 and
they live forever after -- or until the garbage collector kicks in.
Triggering it manually, for demonstration purposes:
>>> import gc; gc.collect()
A says bye!
B says bye!
4
That's for Python 3, in Python 2 the output will be
>>> import gc; gc.collect()
4
The objects will still be released, it's only that you don't see it because
their __del__() methods aren't invoked.
--
https://mail.python.org/mailman/listinfo/python-list
