> The bigger picture here is that catalog changes are supposed to be > executed by C code in response to DDL commands, and it's the C code > that is charged with maintaining catalog consistency. Constraints > would be useful if we supported updating the catalogs with direct > SQL manipulations; but we don't really. You can do that, if you're > a superuser who's willing to take risks, but our policy is that if > you break the catalogs that way you get to keep both pieces. Ah! That makes sense. Constraints are there to help you when inserting rows and when it comes to the catalogs you shouldn't be doing that directly. I think for the diagram I'll add the links manually with the caveat that they are not constraints, just edges in a graph. Thanks for the details Tom! -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
