Joe Conway <mail(at)joeconway(dot)com> writes:
> Plug RLS related information leak in pg_stats view.
BTW, a note about this: I see you bumped catversion in both branches,
which is good, but you bumped it to the same thing, which I think is
bad. Two branches with different initial catalog contents should
not have identical catversions; that more or less defeats the point
of having the version number.
What I've done in such cases is to assign yyyymmdd1 to the back
branch and yyyymmdd2 to HEAD.
regards, tom lane