Re: GiST rtree logic is not right

"John Hansen" <john@geeknet.com.au> writes:
>> I'll look at problem after GiST concurrency. Fixing 
>> rtree_gist is bug a fix, not a new feature, so I'm not 
>> limited by 1 July.

> Wont fixing rtree(_gist) require initdb, since the behaviour of the
> operators will change?

Possibly, but we never guarantee no initdb until final release anyway.

Teodor is right that this is a bug fix and so can be postponed on its
own terms.  But moving rtree_gist into the core looks like a feature
change to me, so if that's going to happen it has to happen before
1 July.  It would be a lot easier to sell that if it gave the right
answers ;-)
        regards, tom lane