Thread

Re: FOR PORTION OF vs. object_aclcheck

Robert Haas <robertmhaas@gmail.com> — 2026-05-13T12:13:09Z

On Tue, May 12, 2026 at 11:37 AM Peter Eisentraut <peter@eisentraut.org> wrote:
> If we assume the type support function approach, I think we don't
> normally check permissions on type support functions.  It is implied
> that if you make use of a type, then the type support functions could be
> executed at any time without further permission checking.

I'm not totally sure we've made the right policy choice there, and I'm
also not sure that we're entirely consistent, but there's certainly
precedent for that approach.

> Maybe we should at least expand the comment to remind future developers
> of this concern, though?

Yeah, if we're not going to add a check, we should definitely write
something in the comment.

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