Pavel Raiskup <praiskup@redhat.com> writes:
> On Monday, July 9, 2018 7:41:59 PM CEST Tom Lane wrote:
>> Hi Pavel! For patches that purport to resolve bugs, we usually like to
>> add a regression test case that demonstrates the bug in unpatched code.
>> Can you provide a small test case that does so? (The BZ you pointed to
>> doesn't seem to address this...)
> Turns out the problem is only related to bit/bit varying type (so the
> patch comments need to be reworded properly, at least) since those are the
> only types which have implemented the f_l2n() callback.
What I'm failing to wrap my head around is why this code exists at all.
AFAICS, gbt_bit_xfrm just forces the bitstring to be zero-padded out to
an INTALIGN boundary, which it wouldn't necessarily be otherwise. But
why bother? That should have no effect on the behavior of bit_cmp().
So I'm speculating that the reason nobody has noticed a problem is that
there is no problem because this code is useless and could be ripped out.
It would be easier to figure this out if the btree_gist code weren't
so desperately undocumented. Teodor, do you remember why it's like
this?
regards, tom lane