The "DBConnId" is very strange ("None" rather than "DB_Connection").Somehow, the "tcache" and "cache" variables are inconsistent. I do not know however, how this can happen. A theoretical possibility would be two different "tcache" entries tcache[i1] = tcache[i2] = query. Then "cache[query]" would be deleted for the first entry and the second entry would result in a "KeyError". However, such a case should be impossible...
Interesting issue. I think the simplest thing that can be done is to add some logging into _cached_result. It should allow to track cache state and operations done on it just before the key error. -- Maciej Wisniowski _______________________________________________ Zope-DB mailing list [email protected] http://mail.zope.org/mailman/listinfo/zope-db
