[ZODB-Dev] Re: ZODB-Dev Digest, Vol 27, Issue 3

Michel Pelletier michel at dialnetwork.com
Fri Jun 3 13:09:43 EDT 2005

> > If you do this from Zope 2.8, though, it picks up Zope 2.8's implementation,
> > and doesn't give a warning.  In short, ExtensionClass is a Zope2 thing, not
> > a ZODB thing (nor even a Zope3 thing).
> Yep, and there was much rejoicing! Sadly, I don't think Zope3 has
> Catalog yet, does it?

Yes 3.1 has a very nice catalog implementation that should be familiar
enough to a ZCatalog user.  The Zope 3 version is much cleaner.  I would
like to see something like Dieter's AdvancedQuery ported to it.

> Zope2 relies on ExtensionClass for the __of__ mechanism to enable
> Acquisition. I realize the implicit acquisition is currently viewed as
> less than ideal, but it seems like Zope3 would still need __of__. That
> must be coming from somewhere else.

Honestly I haven't seen a Zope 3 class yet that subclasses from
Acquisition, and if there is one it's deep in the bowels somewhere where
I don't have to worry about it, and there is no Zope 2 style implicit
acquisition on every single object in the containment hierarchy.  I look
forward to writing a lot less of __of__.

> Actually, ExtensionClass was very easy. As of now, I have a running,
> but poorly tested, standalone Catalog package. (With a distutils setup
> and everything :)  Once I've kicked it around a bit more, I'm going to
> release it, since I haven't seen anyone else actually package it up.

If all you're using is ZODB and a catalog then absolutely I'd recommend
using the Zope 3 version.


> Kevin

More information about the ZODB-Dev mailing list