Rewrite new_persistent_id() to make it clear what's going on.

Reformat all the code to use Python style.

We got an undiagnosed bug report for coptimization, which lead to an
attempt at code review.

Break up a the large new_persistent_id() function into several helpers
-- get_class(), get_class_tuple(), and set_oid().  Each function now
does a little less.  Don't reuse variable names for random purposes.

src/ZODB/Connection.py

@@ -13,7 +13,7 @@
 ##############################################################################
 """Database connection support
 
-$Id: Connection.py,v 1.75 2002/09/07 17:22:09 jeremy Exp $"""
+$Id: Connection.py,v 1.76 2002/09/16 23:50:39 jeremy Exp $"""
 
 from cPickleCache import PickleCache, MUCH_RING_CHECKING
 from POSException import ConflictError, ReadConflictError
@@ -350,7 +350,7 @@ class Connection(ExportImport.ExportImport):
         file=StringIO()
         seek=file.seek
         pickler=Pickler(file,1)
-        pickler.persistent_id=new_persistent_id(self, stack.append)
+        pickler.persistent_id=new_persistent_id(self, stack)
         dbstore=self._storage.store
         file=file.getvalue
         cache=self._cache