Re: rbtree code breaks GIN's adherence to maintenance_work_mem

Robert Haas <robertmhaas@gmail.com> writes:
> On Sat, Jul 31, 2010 at 12:32 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> So it'd definitely be a lot better than now. Maybe there'd be some
>> remaining performance gap for non-pathological cases, but I think we
>> would be willing to pay that in order to avoid bad behavior for the
>> pathological cases. It's difficult to say for sure of course
>> without going to the trouble of coding and testing it.

> Well, it sounds like a reasonable thing to try, then.  You going to
> take a crack at it?

Yeah, I'll have at it.

			regards, tom lane