Jon Nelson <jnelson+pgsql(at)jamponi(dot)net> writes:
> Does this suggest that the config items "disable_seqscan" (and friends)
> should be renamed to "avoid_seqscan" ?
Some of them actually are disable, some are only avoid. The
documentation says which is which, but I don't think creating a
distinction at the naming level would be helpful (especially since
the planner's behavior has been known to change from time to time).
regards, tom lane