Re: Strange behavior with leap dates and centuries BC

--On Montag, Februar 25, 2008 12:00:05 -0500 Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
wrote:

> regression=# select '0001-02-28 BC'::date + 1;
> ?column?
> ---------------
> 0001-02-29 BC
> (1 row)
>
> regression=# select '0002-02-28 BC'::date + 1;
> ?column?
> ---------------
> 0002-03-01 BC
> (1 row)

I stepped through the code in datetime.c and it seems the culprit here is
DecodeDate(). It get's the date string from DecodeDateTime(), but without
the 'BC' century notation. However, it then performs the following check

/* there is no year zero in AD/BC notation; i.e. "1 BC" == year 0 */
if (bc)
{
if (tm->tm_year > 0)
tm->tm_year = -(tm->tm_year - 1);
else
ereport(ERROR,
(errcode(ERRCODE_INVALID_DATETIME_FORMAT),
errmsg("inconsistent use of year %04d and \"BC\"",
tm->tm_year)));
}

bc never becames true during parsing and the final check for the leap date
fails:

/* We don't want to hint about DateStyle for Feb 29 */
if (tm->tm_mday > day_tab[isleap(tm->tm_year)][tm->tm_mon - 1])
{
return DTERR_FIELD_OVERFLOW;
}

Maybe that helps a little bit.

--
Thanks

Bernd