Jeff Davis <pgsql(at)j-davis(dot)com> writes:
> One thing that still doesn't make sense to me is that texteq() is
> bitwise-equality even in 8.3.
Historical artifact ... we made the semantics change some time ago, but
the ensuing change to remove ~=~ didn't happen until 8.4.
> It sounds like Reece Hart can avoid the extra index by making a new
> opclass that's like text_pattern_ops except "~=~" should be named "=".
Unfortunately not, because the LIKE index optimizations are keyed off
the built-in opclasses.
regards, tom lane