This is Florent Guillaume's branch, giving transaction objects
a new beforeCommitHook() method, as proposed by Jim Fulton on
zodb-dev.  Some changes were made to the branch code (comments,
more tests, more words in the docs).

This doesn't have a lot of visible effect on standalone ZODB use, but
is important for Zope 2.8.

Here are checkin msgs from the branch relating to ZODB code:

    r29872 | jim | 2005-04-04 07:04:39 -0400 (Mon, 04 Apr 2005) | 3 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/ZODB/serialize.py

    For instances of persistent classes, save a class-module/clas-name
    tuple if the class has a non-empty module string.

    ------------------------------------------------------------------------
    r29871 | jim | 2005-04-04 07:04:33 -0400 (Mon, 04 Apr 2005) | 6 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/ZODB/Connection.py

    Changed sub-transaction abort code to not invalidate created objects.
    There's really no point, because created objects will be unreachable
    after invalidating old objects.  Also, if a created object is
    non-ghostifiable, it will try to load it's state again, and then
    either it would fail or it would load non-committed state.

    ------------------------------------------------------------------------
    r29870 | jim | 2005-04-04 07:04:27 -0400 (Mon, 04 Apr 2005) | 3 lines
    Changed paths:
       A /Zope/branches/jim-fix-zclasses/lib/python/ZODB/persistentclass.py
       A /Zope/branches/jim-fix-zclasses/lib/python/ZODB/persistentclass.txt
       A /Zope/branches/jim-fix-zclasses/lib/python/ZODB/tests/testpersistentclass.py

    Added ZClass-independent test of (and possible base class for)
    persistent-class support machinery.

    ------------------------------------------------------------------------
    r29867 | jim | 2005-04-04 07:03:48 -0400 (Mon, 04 Apr 2005) | 2 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/transaction/_manager.py

    Added missing arguments to commit and abort.

    ------------------------------------------------------------------------
    r29774 | jim | 2005-04-01 06:24:27 -0500 (Fri, 01 Apr 2005) | 3 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/transaction/__init__.py

    Changed to use methods of a threaded manager directly, rather than
    through wrapper functions.

    ------------------------------------------------------------------------
    r29773 | jim | 2005-04-01 06:24:25 -0500 (Fri, 01 Apr 2005) | 3 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/transaction/_manager.py

    Added commit and abort methods to transaction managers.
    This makes direcr use of managers simpler.

    ------------------------------------------------------------------------
    r29101 | jim | 2005-02-10 07:44:52 -0500 (Thu, 10 Feb 2005) | 2 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/ZODB/Connection.py

    Removed an unused attribute.

    ------------------------------------------------------------------------
    r29100 | jim | 2005-02-10 07:42:35 -0500 (Thu, 10 Feb 2005) | 3 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/transaction/_transaction.py

    Changed an XXX comment to an ordinary comment, explaining the relevent
    issue.

    ------------------------------------------------------------------------
    r29099 | jim | 2005-02-10 07:41:58 -0500 (Thu, 10 Feb 2005) | 3 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/ZODB/Connection.py

    Changed some XXX comments to ordinary comments explaining the relevent
    issues.

    ------------------------------------------------------------------------
    r29071 | jim | 2005-02-07 07:36:05 -0500 (Mon, 07 Feb 2005) | 3 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/transaction/_transaction.py

    Added a sanity check to avoid registration of objects without a
    manager. (Perhaps this should be an assert.)

    ------------------------------------------------------------------------
    r29067 | jim | 2005-02-07 07:35:56 -0500 (Mon, 07 Feb 2005) | 15 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/ZODB/Connection.py

    Changed the strategy for handling invalidation of classes.
    No-longer use setklassstate.  Instead, just call _p_invalidate, as
    with any other object.  This changes didn't break any tests, so I
    assume that this was untested. :(

    Changed the strategy for invalidating objects.  Non-ghostifiable
    objects will load their state when they are invalidated.  We have to
    worry about other invalidations that come in while this is happening.
    See the comment in _flush_invalidations.

    We need to force all invalidations to take this into account, by going
    through _flush_invalidations.  I haven't done this yet, but I have
    left some XXX comments in places where it needs to be done to remind
    myself that this needs to be done.

    ------------------------------------------------------------------------
    r29066 | jim | 2005-02-07 07:35:54 -0500 (Mon, 07 Feb 2005) | 8 lines
    Changed paths:
       M /Zope/branches/jim-fix-zclasses/lib/python/persistent/cPickleCache.c

    Changed the strategy for handling invalidation of classes.
    No-longer use setklassstate.  Instead, just call _p_invalidate, as
    with any other object.  This changes didn't break any tests, so I
    assume that this was untested. :(

    Change invalidation to not swallow errors. (Swallowing errors here was a
    travesty!)

Port from Zope 2.7 branch.

The various flavors of the ``check2ZODBThreads`` and ``check7ZODBThreads``
tests are much less likely to suffer sproadic failures now.

Since ZODB needs to work with Zope trunk, keeping this dependence
in synch is important.

Zope trunk gets it.

development get a release tag to use.

The Persistence module no longer exists in Zope3's idea of what
ZODB is, but this test relies on it.  Make it tolerate the
absence of Persistence.

(backported from trunk revisions 29818 and 29819)

Noet that the URL doesn't exist yet.

The ZODB trunk is for 3.5 now.

This is news about ZEO protocol changes.  The promised
update of README.txt will have to be rewritten from
scratch for 3.4 (the new text in 3.3's README.txt isn't
quite right for the 3.4 release).

Rewrite ZEO protocol negotiation.

3.3 should have bumped the ZEO protocol number (new methods were
added for MVCC support), but didn't.  Untangling this is a mess.

the most common cases (commit() and abort()) -- it's not really
obvious that shortcut spelling exist for those.

Merge checkin made from a wrong project.

Believe it or not, ZopeUndo is ZEO code, not Zope code.

r29726 | slinkp | 2005-03-30 02:11:22 -0500 (Wed, 30 Mar 2005) | 3 lines
Changed paths:
   M /Zope/trunk/lib/python/ZopeUndo/Prefix.py
   M /Zope/trunk/lib/python/ZopeUndo/tests/testPrefix.py

Merged slinkp_1726_zopeundo: avoid showing undo transactions in
the wrong folder.

This always slept for 3 seconds, waiting for the storage to
disconnect.  This loses on two counts:

1. Since the timeout is set to 1 second, it typically sleeps
   longer than necessary.

2. Since there's no predicting thread and process scheduling,
   3 seconds isn't always long enough, and rare failures have
   been reported against this test.

Instead we do a polling loop now.  This typically succeeds in
a little more than a second, speeding the normal case.  The
loop will continue trying for up to a minute if not.

Change default port # from 9999 to 8100, to match zope.conf
(well, on Zope trunk anyway -- the ZEO client example appears
to have gone missing from zope.conf on Zope3 trunk).

s/zeo/ZEO/ in logger name, to match other ZEO logger names.

Collector #1734.  Critical bug in BTree conflict resolution.

Stop silent data loss in some BTree cases where a transaction adds
a new key to a bucket while a concurrent transaction deletes all
keys from the same bucket.

Still needs porting to 3.2 line.

Various Zopes probably need changes too (to avoid the new deprecation
warnings).

Jim (Fulton) points out that it's not entirely clear that
the new format can't be read by older ZODBs, and Tim doesn't
think it's worth the bother to find out for sure.

Added news blurbs about fsIndex improvements, and about that
.index files written by 3.4 won't be readable by earlier ZODBs.

test.py ran these tests ran after "setup build_ext -i", but
not after "setup build".  Repaired by teaching setup.py to
copy more stuff out of the relevant directories.

There's a mystery:  the tests for it appear to run after
"build_ext -i", but not after "build".  But that appears
to be true for the long-stitched-in zope.interface too.

This introduces a "multi-database" concept (a simplification
of Jim's Wiki proposal), and adds many interface definitions.

Work done during the PyCon 2005 ZODB sprint, by Christian
Theune, Jim Fulton and Tim Peters.