Mario Splivalo <mario.splivalo@megafon.hr> writes:
> Huh! As I was explained od pg-sql list by Tom Lane, this should not
> happen, so it is a bug! :) I apologize for the inconvenience.
It's already fixed:
http://archives.postgresql.org/pgsql-committers/2010-04/msg00126.php
regards, tom lane