Re: Bug #6593, extensions, and proposed new patch policy

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> On Wed, Apr 18, 2012 at 5:27 PM, Alvaro Herrera
> <alvherre(at)commandprompt(dot)com> wrote:
>> Here's a patch for that.

> Looks sane on a quick once-over. I do wonder about the comment,
> though. If we add ALTER EXTENSION .. OWNER, should that try to change
> the ownership of the objects contained inside the extension?

I would certainly think that not doing so would violate the principle
of least astonishment.

> Your
> comment implies that the answer should be yes, but I'm not totally
> convinced... what if the user has altered the ownership of the
> objects manually, for example?

So? ALTER OWNER doesn't care about the previous ownership of objects,
it just reassigns them as told. So even if that had been done, I'd
expect the post-ALTER state to be that everything has the new owner.

However, ignoring that issue for the moment, this patch is making me
uncomfortable. I have a vague recollection that we deliberately omitted
ALTER EXTENSION OWNER because of security or definitional worries.
(Dimitri, does that ring any bells?) I wonder whether we should insist
that the new owner be a superuser, as the original owner must have been.

regards, tom lane