Abhijit Menon-Sen <ams@2ndQuadrant.com> writes:
> I don't think it's fair to mark this as returned with feedback without a
> more detailed review (I think of returned with feedback as providing a
> concrete direction to follow). I've set it back to "needs review".
> Does anyone else want to look at this patch?
I offered a few comments. I think it might also be useful to try to
quantify the worst-case performance cost for this, which would arise
for a single-table query on a table with lots of indexes. Don't
have time for that myself though.
regards, tom lane