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

On Thu, Jun 2, 2022 at 1:17 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Point 2 would cause every existing pg_dumpall script to fail, which
> seems like kind of a large gotcha. Less unpleasant alternatives
> could include
>
> * Continue to accept the syntax, but ignore it, maybe with a WARNING
> for the alternative that doesn't correspond to the new behavior.
>
> * Keep pg_authid.rolinherit, and have it act as supplying the default
> behavior for subsequent GRANTs to that role.

Of those two alternatives, I would certainly prefer the first, because
the second doesn't actually get rid of the ugly wart. It just adds a
non-ugly thing that we have to maintain along with the ugly thing,
apparently in perpetuity. If we do the first of these, we can probably
remove the obsolete syntax at some point in the distant future, and in
the meantime, we don't have to figure out how it's supposed to
interact with existing features or new ones, since the actual feature
is already removed.

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