Re: BUG #19040: Memory leak in hashed subplan node due to missing hashtempcxt reset

feichanghong <feichanghong@qq.com> writes:
> Thanks for the analysis! v04-0002 looks fine to me — please just apply it on
> HEAD.

Done that way.

I thought a bit about your suggestion of adding some kind of
assertion check for memory leaks, but it looks too messy and
specialized as-is.  Maybe with reflection we can find a more
generic idea.

            regards, tom lane