[Checkins] SVN: zc.intid/trunk/README.txt clarify the differences

Fred Drake fdrake at gmail.com
Fri Nov 20 10:45:27 EST 2009


Log message for revision 105910:
  clarify the differences
  

Changed:
  U   zc.intid/trunk/README.txt

-=-
Modified: zc.intid/trunk/README.txt
===================================================================
--- zc.intid/trunk/README.txt	2009-11-20 14:54:18 UTC (rev 105909)
+++ zc.intid/trunk/README.txt	2009-11-20 15:45:26 UTC (rev 105910)
@@ -8,12 +8,23 @@
 undesirable, such as in search indices or any code that needs an easy
 hash of an object.
 
-This is similar to the ``zope.intid`` package, but has fewer
-dependencies and induces fewer conflict errors, since object ids are not
-used as part of the stored data.  The id for an object is stored in an
-attribute of the object itself, with the attribute name being configured
-by the construction of the id utility.
+This is similar to the ``zope.intid`` package, but with two distinct
+advantages:
 
+- Induces fewer conflict errors, since object ids are not used as part
+  of the stored data.  The id for an object is stored in an attribute of
+  the object itself, with the attribute name being configured by the
+  construction of the id utility.
+
+  This does require that the object "play nice" with this approach.  At
+  a minimum, the attributes used to store ids on objects should
+
+  - persist with the rest of the object's state, and
+
+  - not be modified by the object.
+
+- Fewer dependencies.
+
 Events generated on the assignment and removal of ids are generated by
 the ``register`` and ``unregister`` methods rather than by the callers
 of those methods.



More information about the checkins mailing list