Bruce Momjian <pgman@candle.pha.pa.us> writes:
> If I remember correctly, you didn't like the index routines reading the
> tuple information, or something like that, but there was a performance
> benefit for duplicate keys, so I think we should re-investigate this.
I don't see the actual patch either in the hackers or patches archives,
nor on your to-apply pages, making it a bit difficult to re-investigate.
Where was it posted anyway?
regards, tom lane