On Fri, April 6, 2007 16:18, John Mudd wrote: > Is there a similar risk with m_home?
No, that's a different matter, I think. Invoking a prepared statement is something you do entirely within the scope of one transaction. It should be obvious that you can't start invoking a statement in the context of one transaction, then not only close that transaction but even destroy its representing object, and then continue your invocation. That'd be as silly as destroying a connection and then continuing to use your transaction. Jeroen _______________________________________________ Libpqxx-general mailing list Libpqxx-general@gborg.postgresql.org http://gborg.postgresql.org/mailman/listinfo/libpqxx-general