> >> While we don't _need_ it, it would make our system more consistent; we
> >> have made similar changes for views in other areas.
> >
> > I'm not sure it'd make the system more consistent. Because the SQL
> > standard says you use GRANT ON TABLE for a view. we'd have to always
> > ensure that we accepted that; whereas in at least some other places
> > we are trying to be picky about TABLE/VIEW/SEQUENCE actually matching
> > the object type.
> >
> > Given the spec precedent, I'm inclined to leave it alone. It's not like
> > there aren't plenty of other SQL quirks that surprise novices.
>
> I fail to understand why it's advantageous to artificially create
> surprising behavior. There are cases where PostgreSQL now accepts
> either ALTER VIEW or ALTER TABLE where it previously accepted only the
> latter, so the situation is hardly without precedent. I find it
> exceedingly unlikely that anyone is relying on GRANT ON VIEW to NOT
> work.
Yes, I assumed we were following the recent work on ALTER TABLE/VIEW
with GRANT/REVOKE. Peter, Tom, how is GRANT/REVOKE different?
--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ If your life is a hard drive, Christ can be your backup. +
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
No comments:
Post a Comment