On Mon, Oct 16, 2023 at 11:50:08AM +1300, Thomas Munro wrote:
> On Mon, Oct 16, 2023 at 11:24 AM Thomas Munro <thomas.munro@gmail.com> wrote:
>> On Mon, Oct 16, 2023 at 10:57 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
>>> I'm tempted to wonder if this helps:
>>>
>>> - result->time = t->time + (t->zone - tz) * USECS_PER_SEC;
>>> + result->time = t->time + (int64) (t->zone - tz) * USECS_PER_SEC;
All that should use TZNAME_FIXED_OFFSET as timezone type, and I don't
really see why this would overflow..
Perhaps a more aggressive (int64) ((t->zone - (int64) tz) *
USECS_PER_SEC) would help?
>> I wanted to be able to try this and any other theories and managed to
>> build the tip of master on cfarm111 with the same CC and CFLAGS as
>> Noah used, but the problem didn't reproduce! Hmm, I didn't enable any
>> extra options, so now I'm wondering if something in some random header
>> somewhere is involved here... trying again with more stuff turned
>> on...
>
> Oh, I can't use any of the handrolled packages in ~nm due to
> permissions. I tried enabling perl from /opt/freeware (perl is my
> usual first guess for who is !@#$ing with the system headers), but the
> test passes.
Another theory would be one of these weird compiler optimization issue
from xlc? In recent history, there was 8d2a01ae12cd.
--
Michael