James Coleman <jtc...@gmail.com> writes: > On Tue, Jun 25, 2024 at 4:59 PM Tom Lane <t...@sss.pgh.pa.us> wrote: >> Uh ... what? I recall something about that with respect to certain >> features such as nextval(), but you're making it sound like there >> is something generic going on with DEFAULT.
> Hmm, I guess I'd never considered anything besides cases like > nextval() and now(), but I see now that now() must also be special > cased (when quoted) since 'date_trunc(day, now())'::timestamp doesn't > work but 'now()'::timestamp does. Hmm, both of those behaviors are documented, but not in the same place and possibly not anywhere near where you looked for info about DEFAULT. For instance, the Tip at the bottom of section 9.9.5 https://www.postgresql.org/docs/current/functions-datetime.html#FUNCTIONS-DATETIME-CURRENT explains about how 'now'::timestamp isn't what to use in DEFAULT. regards, tom lane
