Re: replacing role-level NOINHERIT with a grant-level option

Поиск
Список
Период
Сортировка
От Robert Haas
Тема Re: replacing role-level NOINHERIT with a grant-level option
Дата
Msg-id CA+TgmoZ90mLF6qogXuoyQYVoO7UcRfr1Mdnp=O+W10gNx_iObA@mail.gmail.com
обсуждение исходный текст
Ответ на Re: replacing role-level NOINHERIT with a grant-level option  (Nathan Bossart <nathandbossart@gmail.com>)
Ответы Re: replacing role-level NOINHERIT with a grant-level option  (Robert Haas <robertmhaas@gmail.com>)
Список pgsql-hackers
On Sun, Jul 3, 2022 at 1:17 PM Nathan Bossart <nathandbossart@gmail.com> wrote:
> If by "bolder" you mean "mark [NO]INHERIT as deprecated-and-to-be-removed
> and begin emitting WARNINGs when it and WITH INHERIT DEFAULT are used," I
> think it's worth consideration.  I suspect it will be hard to sell removing
> [NO]INHERIT in v16 because it would introduce a compatibility break without
> giving users much time to migrate.  I could be wrong, though.

It's a fair point. But, if our goal for v16 is to do something that
could lead to an eventual deprecation of [NO]INHERIT, I still think
removing WITH INHERIT DEFAULT from the patch set is probably a good
idea. Perhaps then we could document that we recommend using the
grant-level option rather than setting NOINHERIT on the role, and if
users were to follow that advice, eventually no one would be relying
on the role-level property anymore. It might be optimistic to assume
that users will read the documentation, let alone follow it, but it's
something. Now, that does mean accepting a compatibility break now, in
that flipping the role-level setting would no longer affect existing
grants, but it's less of a compatibility break than I proposed
originally, so maybe it's acceptable.

-- 
Robert Haas
EDB: http://www.enterprisedb.com



В списке pgsql-hackers по дате отправления:

Предыдущее
От: Alvaro Herrera
Дата:
Сообщение: Re: doc: BRIN indexes and autosummarize
Следующее
От: Amit Kapila
Дата:
Сообщение: Re: logical replication restrictions