Tomas Vondra <tomas.vondra@2ndquadrant.com> writes:
> Anyway, I think you're right we're going in circles here. I think we
> both presented all the arguments we had and we still disagree. I'm not
> going to continue with this - I'm unlikely to win an argument against
> two committers if that didn't happen until now. Thanks for the
> discussion though.
Since we're hard up against the 9.5beta1 deadline, I've made an executive
decision to commit just the minimal change, which I view as being to
constrain the array size to MaxAllocSize where it has been all along.
I found a second rather serious bug in the new hash code while doing that
--- it failed to ensure nbuckets was actually a power of 2 --- which did
not improve my opinion of it one bit.
It's clear from this discussion that there's room for further improvement
in the hashtable sizing behavior, but I think that's 9.6 material at
this point.
regards, tom lane