Michael Paquier <michael@paquier.xyz> writes:
> On Thu, Jan 01, 2026 at 10:00:00AM +0200, Alexander Lakhin wrote:
>> replace_s(..., int * adjustment) -> replace_s(..., int * adjptr)
> Updating this one contradicts with the upstream sources, see
> b464e51ab32f that has done the opposite switch.
It looks like upstream has fixed the problem -- not by making the
two instances match, but by removing the parameter altogether
(see their recent commit 16edd5abb9ce7386e9208b998cd43524480cb801).
Maybe it's time for another sync with upstream? But if we don't
feel like doing that right now, I think Alexander's proposed patch
is okay. It won't make the next sync any harder.
regards, tom lane
