[ZODB-Dev] ZODB 3.2.4 strange getattr, hasattr behaviour

[Syver Enstad]

"Tim Peters" <tim at zope.com> writes:


> The question I have is about "add one more occurrence ...":  if this is
> something that routinely bites you, why do you create class data attributes?

I don't usually, but I tried it for a quick fix for a new attribute in
a Persistent class, since __getstate__ updating doesn't work in ZODB
(the _p_changed and so on). Actually it has only been one time
previous but that time confused me a lot too.

I did something like this (a couple of years ago):

class TheClass:
    _attr = None
    def myMethod(self):
        if self._attr is None:
            self._attr = SoAndSo

inst = TheClass()
inst.myMethod()
assert TheClass._attr == SoAndSo # this fails as I guess most of you know

So the case with my upgrade script just highlights that once more.
  
> Unlike class method attributes, class data attributes are almost never
> necessary.  IOW, if punching yourself in the eye hurts, there are
> alternatives to wishing your knuckles were softer.  For example, you could
> wish that your eye were harder <wink>.

The first thing I did after discovering what I had done was to get rid
of them by using an upgrade script instead.