Mark Knox <segfault@hardline.org> writes:
> Well, this patch seems to produce attlens of 6 as desired, but it
> causes many (13) of the regression tests to fail. Do you want to see
> the regression.diffs?
>>
>> Please.
> See attached.
Does look pretty broken, but I don't see how my idea would have led to
all this other stuff failing. Anyway, I guess the path of least
resistance is to install your ARM-specific packing patch. It's
important to make sure that sizeof(ItemPointerData) is 6 if at all
possible, since it will cost you four or so wasted bytes in every
tuple header if it's not. Will take care of it for RC2.
regards, tom lane