Andrey Borodin <amborodi...@gmail.com> writes: > On Wed, Feb 15, 2023 at 7:08 AM Nikolai <pgni...@gmail.com> wrote: >> The patch attached simply throws an error when an overflow is >> detected. However I'm not sure this is a reasonable approach for a >> code path that could be very hot in some workloads.
> Given the extraordinary amount of overflow checks in the nearby code > of timestamp.c, I'd say that this case should not be an exception. Yeah, I don't think this would create a performance problem, at least not if you're using a compiler that implements pg_sub_s64_overflow reasonably. (And if you're not, and this bugs you, the answer is to get a better compiler.) > By chance did you look at all other nearby cases, is it the only place > with overflow? That was my immediate reaction as well. regards, tom lane