Guido van Rossum wrote:
> I don't see a way that __del__ could be invoked and access the NULL
> between the Py_CLEAR() call and the Py_INCREF() call in the second
> version.
Consider this:
class Bad:
def __del__(self):
obj.some_method()
obj.set_foo(Bad())
obj.set_foo(None)
Now, the second set_foo does a Py_CLEAR on foo, which invokes
Bad.__del__, which invokes a method on obj, which crashes as foo
is NULL - it has not yet been assigned Py_None as expected.
Regards,
Martin
_______________________________________________
Python-Dev mailing list
Python-Dev@python.org
http://mail.python.org/mailman/listinfo/python-dev
Unsubscribe:
http://mail.python.org/mailman/options/python-dev/archive%40mail-archive.com
