Robert Haas <> writes:
> Oh, one other thing about this: I'm not going to try to defend
> whatever confusion between subplans and subqueries appears in that
> comment, but prior to upper planner pathification I think we were dead
> in the water here anyway, because the subquery was going to output a
> finished plan, not a list of paths.  Since finished plans aren't
> labeled as to parallel-safety, we'd have to conservatively assume that
> the finished plan we got back might not be parallel-safe.  Now that
> we're using the path representation throughout, we can do better.

Well, if that were still the state of affairs, we could certainly consider
adding a parallel_safe flag to Plans as well as Paths.  But as you say,
it's moot now.

                        regards, tom lane

Sent via pgsql-hackers mailing list (
To make changes to your subscription:

Reply via email to