Thomas Munro <thomas(dot)munro(at)enterprisedb(dot)com> writes:
> As noted in a nearby review of some similar code[1], commit 7a2fe9bd
> made merge append very slightly more efficient, but nobody told
> cost_merge_append about the change. I doubt it makes much difference
> to the final cost in practice but I figured it might be worth
> correcting the comment. Does this make sense?
Agreed, pushed.
regards, tom lane