Re: Remove dependence on integer wrapping

Nathan Bossart <nathandbossart(at)gmail(dot)com> writes:
> On Sun, Jun 09, 2024 at 09:57:54PM -0400, Tom Lane wrote:
>> I think you could replace the whole thing by using overflow-aware
>> multiplication and addition primitives in the result calculation.

> I was still confused by the comment about 1999, but I tracked it down to
> commit 542eeba [0]. IIUC it literally means that we need special handling
> for that date because POSTGRES_EPOCH_JDATE is 2000-01-01.

Yeah, I think so, and I think we probably don't need any special care
if we switch to direct tests of overflow-aware primitives. (Though
it'd be worth checking that '1999-12-31 24:00:00'::timestamp still
works. It doesn't look like I actually added a test case for that.)

regards, tom lane