> -----Original Message----- > From: Merlin Moncure [mailto:mmoncure@xxxxxxxxx] > Sent: Thursday, December 20, 2007 8:30 AM > To: Roberts, Jon > Cc: Alvaro Herrera; Trevor Talbot; Joshua D. Drake; Kris Jurka; Jonah H. > Harris; Bill Moran; pgsql-performance@xxxxxxxxxxxxxx > Subject: Re: viewing source code > > On Dec 20, 2007 9:07 AM, Roberts, Jon <Jon.Roberts@xxxxxxxxxxx> wrote: > > So your suggestion is first to come up with a query that dynamically > checks > > permissions and create a view for it. Secondly, change pgAdmin to > reference > > this view in place of pg_proc. Actually, it should be extended to all > > This solution will not work. It requires cooperation from pgAdmin > which is not going to happen and does nothing about psql or direct > queries from within pgadmin. Considered from a security/obfuscation > perspective, its completely ineffective. As I've said many times, > there are only two solutions to this problem: > > 1. disable permissions to pg_proc and deal with the side effects > (mainly, pgadmin being broken). > > 2. wrap procedure languages in encrypted handler (pl/pgsql_s) so that > the procedure code is encrypted in pg_proc. this is an ideal > solution, but the most work. > I think there is an option 3. Enhance the db to have this feature built in which is more inline with commercial databases. This feature would drive adoption of PostgreSQL. It isn't feasible in most companies to allow everyone with access to the database to view all code written by anyone and everyone. For instance, you could have a Finance group writing functions to calculate your financial earnings. These calculations could be changing frequently and should only be visible to a small group of people. If the calculations were visible by anyone with database access, they could figure out earnings prior to the release and thus have inside information on the stock. Jon ---------------------------(end of broadcast)--------------------------- TIP 2: Don't 'kill -9' the postmaster
