XXX This creates a new requirement: If a store implements restore(),
it's iterator() implementation must support the data_txn option.

XXX This version fails silently when it should be able to write a
backpointer, but can't.

Some storages get confused by tid == 0.

The Record() object now has a data_txn attribute that is either None
or the id of the transaction that contains the data used by the
current record.  Example: When transactionalUndo() modifies an object,
it typical creates a new data record that points at the transaction
before the undo.  The new record contains the same logical data as the
record it refers to.  (For consistency purposes, this is a stronger
claim than that the pickles in two different data records are the
same.)

Add a test of the new iterator() feature in TransactionalUndoStorage.

Share backpointer implementation between _loadBack() and
_loadBackPOS() to avoid duplicate code.

Simplify interface between FileIterator and RecordIterator.  Don't
pass unused arguments and don't pass nested tuples.

Use global variables to hold trans and data struct formats.

Reformat several docstrings.

Add whitespace to RecordIterator so that I can read it.

This change is an untested improvement.  If some object is in the
index but unloadable for some reason, report it along with the
traceback.  When possible, report what objects refer to the
un-loadable one.

dangle.py script that creates one.

It appears the initialization was accidentally removed in rev. 1.51.
Add comment about why percachedel is not initialized here.

supported.

passed around and revised for almost a month.)  This is accurate, but
not necessarily complete.

by assuming all pids in the pidfile must be killed.

failure of chdir.

signal meanings: rotate_logs_handler() was changing the SIGHUP handler
to rotate the logs!  zdaemon did not do anything wrong. :-)
Fixed by *removing* the signal() call inside the handler -- this is
not needed in modern Unixes.

Also changed setup_signals(): there was a typo in the name SIFXFSZ(!);
and made all the signal conditional on whether the signal module has
that particular signal (e.g. on Windows, a signal module exists, but
it only defines a few signals).

1. Use SIGHUP for restart instead of SIGINT, and use SIGUSR2 to reopen
   the log file instead of SIGHUP.  This is the same standard as for
   the Zope server.

2. When using zdaemon, let zdaemon write its pid to the zeo_pid file;
   when not using zdaemon, write only our own pid to the zeo_pid
   file.  Never write two pids.

Note that I'm not sure the pids written to the pidfile always make
sense.

documentation.

commit it to the CVS repository.

using the ZConfig.Interpolation module.

in addition to the loads and stores.

This is still in development.

days.  At least the script zeostart.py did this, and it was a clone of
ZEO1/start.py except for that added feature.  Our sysadmin desperately
wants that feature.

records (all the broken records I've seen had a zero timestamp).

Get rid of $Id$ in docstring (it's seen rarely in ZEO).
Append a newline to the pid file.