Robert Haas <robertmhaas@gmail.com> writes:
> OK, I've committed something along those lines. Thanks for the
> advice, and feel free to whack it around if you have an idea for
> improving it further - though IMHO this is good enough for 9.6.
The committed patch looks fine to me. WRT your commit-message comment
about pushing down only parallel-safe tlist expressions, I would strongly
advise leaving that for another day. I have some irons in the fire
concerning ways to push down computation of index expressions (so that
we can consistently get f(x) out of an index on f(x) rather than
recomputing it), and managing parallel-safe expressions seems like it'd
benefit from that infrastucture too.
regards, tom lane
