Re: pg_execute_from_file, patch v10

Itagaki Takahiro <itagaki.takahiro@gmail.com> writes:
> On Tue, Dec 14, 2010 at 12:02, Robert Haas <robertmhaas@gmail.com> wrote:
>> As Tom says, this is clearly not going to fly on security grounds.

> If it's a security hole, lo_import() should be also a hole
> because we can use lo_import() and SELECT * FROM pg_largeobject
> for the same purpose...

lo_import is superuser-only.  If we design this feature so that it will
forever have to be superuser-only, to get a behavior that I think we
don't even *want*, I believe we're making a serious error.
        regards, tom lane