Tom Lane wrote:
> Magnus Hagander <magnus@hagander.net> writes:
>> Tom Lane wrote:
>>> * Just remove the above-quoted lines. Superusers should be allowed to
>>> shoot themselves in the foot. (I'm not actually sure that there would
>>> be any bad consequences from putting an ordinary table into pg_global
>>> anyway.
>
>> Is there ever *any* reason for doing this?
>
> Probably not a good one, and I suspect there would be some funny
> misbehaviors if you were to clone the database containing the table.
> The table would be physically shared but logically not.
yuck.
> What I'm inclined to do about it is is adopt my suggestion #2 (move the
> location of the defense), since "permission denied" for a superuser is
> a pretty unhelpful error message anyway.
Ok. Works for me.
>>> * Decide that we should allow anyone to do pg_tablespace_size('pg_global')
>>> and put in a special wart for that in dbsize.c. This wasn't part of
>>> the original agreement but maybe there's a case to be made for it.
>
>> That's pretty much the same thing, right?
>
> Well, no, I was suggesting that we might want to special-case pg_global
> as a tablespace that anyone (superuser or no) could get the size of.
> This is actually independent of whether we change the aclmask behavior.
Oh, ok, I see. Then my vote is for the other solution = not allowing
anybody to do this.
//Magnus