alvherre@postgresql.org (Alvaro Herrera) writes:
> While at it, mark a couple of items completed in 8.4:
> ! o -Prevent long-lived temporary tables from causing frozen-xid
> advancement starvation
Hum, should we really cross that off as "done"? I know we put in a hack
to clean up temp tables that were orphaned by crashed backends; but what
of the case where a session creates a temp table and then just sits on
it for billions of transactions?
regards, tom lane