Might as well just get rid of that one; Peter's right about the security hole.
The simplest fix I see is to allow SET AUTHORIZATION only in superuser-owned
functions. It would still be potentially useful that way. Doing this the
"right" way (with users needing regrantable privileges, etc.) would involve
too much effort for too little reward, IMHO.
Mark
Bruce Momjian wrote:
>
> I am backing out this SET AUTHORIZATION patch until we can resolve the
> security issues. It will remain in the patch queue at:
>
> http://candle.pha.pa.us/cgi-bin/pgpatches
>