[ZODB-Dev] Tracking down a freeze (deadlock?)

[Jeremy Hylton]

On Wed, 02 Mar 2005 12:19:06 -0500, Jim Fulton <jim at zope.com> wrote:
> Tim Peters wrote:
> ...
> > Why is that?  As far as I'm concerned, ZODB doesn't support persistent
> > objects with __del__ methods -- it was never intended to.
> 
> I'm not aware of any specific decision not to support __del__. I agree
> with Dieter that if we make such a decision, we should go out of our way
> to advertize, and perhaps, enforce it.

I recall talking about years ago.  At the time, I think we didn't want
to support __del__ because it semantics were unclear in the face of
persistence.  But an _p_del() method could be defined to have
semantics that made sense.  Then we were left with the question of
sensible semantics :-).  Possibilities include:

 - on ghostification
 - on actual eviction from cache
 - on deletion from the database (during/after pack)

Still a question about what to do for invalidations, which can be
caused by modification by the current transaction and by another
transaction.

Jeremy