[ZODB-Dev] new-style class status

[Tim Peters]

[John Belmonte]
> I'd like to know what the status of the zodb33-devel-branch is,

It's currently inactive within Zope Corp.  MVCC is a higher priority here.
That in turn has been a lower priority recently than customer (read
"paying") work.

> and to get some idea when a new-style class implementation will at
> least be hobbling well enough for us mortals to use it.

Have you tried zodb33-devel-branch?  We released 3.3a1 from that about 3
months ago:

    http://zope.org/Products/ZODB3.3/

Note that this requires Python 2.3 (under any version of 2.2, it will just
segfault).  IIRC, we got one piece of feedback about it.

I don't what its future may be.  Most likely now seems that MVCC work will
continue to take precedence, get merged into the ZODB3 trunk, and eventually
released as 3.3 (along with scads of bugfixes accumulated on the 3.1 and 3.2
branches).  Merging zodb33-devel-branch into the trunk "should be" the
priority after that, according to me, but I don't set priorities here.

> I'm trying to make my application free of write conflicts, and have a
> lot of mixed operations between scalar amounts that need conflict
> resolution and regular numbers.  I've prototyped a scalar class that
> works like a builtin number and is a Persistent object that can do
> conflict resolution (as long as you stick to operators like +=, etc.),
> but of course extension class comparison operators are broken.
>
> My app is only in development so I don't mind working with an unstable
> ZODB, and may have some time to offer if there is anything I can do to
> get this ZODB improvement moving.

One thing we need is a reimplementation of ExtensionClass as a Python
new-style class.  The current ExtensionClass implementation is much more
complicated than necessary now that Python makes far fewer distinctions
between types and classes, and seems essentially unmaintainable to me.
ExtensionClass in some API-compatible form has to continue to survive, for
compatibility with applications already using ExtensionClass.  I know Jim
(Fulton) looked at this some Friday, but don't know how far he got.