Re: BUG #18240: Undefined behaviour in cash_mul_flt8() and friends
От | Tom Lane |
---|---|
Тема | Re: BUG #18240: Undefined behaviour in cash_mul_flt8() and friends |
Дата | |
Msg-id | 2433182.1702391878@sss.pgh.pa.us обсуждение исходный текст |
Ответ на | Re: BUG #18240: Undefined behaviour in cash_mul_flt8() and friends (David Rowley <dgrowleyml@gmail.com>) |
Ответы |
Re: BUG #18240: Undefined behaviour in cash_mul_flt8() and friends
|
Список | pgsql-bugs |
David Rowley <dgrowleyml@gmail.com> writes: > On Tue, 12 Dec 2023 at 22:03, Michael Paquier <michael@paquier.xyz> wrote: >> And numeric has much better code coverage and support. I am wondering >> whether we've reached the point where it would be better to remove it >> entirely from the tree, and just tell people to use numeric. > We could extract it into a contrib module. Perhaps, but ... > That might help reduce new usages of it and would also allow people > who have large tables using the money type but can't realistically > wait out a table rewrite to upgrade to a newer version of Postgres. ... I doubt that a contrib module would solve the problem for people who can't afford a rewrite. pg_upgrade requires that datatype OIDs stay the same, which is something I don't believe a contrib module could manage. We've run into that before if memory serves. Is it time to do something about that? Perhaps we could allow extension modules to use binary_upgrade_set_next_pg_type_oid, and then somehow reserve the money and _money OIDs forever? regards, tom lane
В списке pgsql-bugs по дате отправления: