Re: [v9.3] Row-Level Security
Kohei KaiGai <kaigai@kaigai.gr.jp>
From: Kohei KaiGai <kaigai@kaigai.gr.jp>
To: David Fetter <david@fetter.org>
Cc: Robert Haas <robertmhaas@gmail.com>, Alvaro Herrera <alvherre@2ndquadrant.com>, Dean Rasheed <dean.a.rasheed@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>, Florian Pflug <fgp@phlo.org>, PgHacker <pgsql-hackers@postgresql.org>
Date: 2012-12-03T15:36:12Z
Lists: pgsql-hackers
2012/12/3 David Fetter <david@fetter.org>: > On Sun, Nov 25, 2012 at 03:20:28PM +0100, Kohei KaiGai wrote: >> > However, UPDATE / DELETE support is not perfect right now. >> > In case when we try to update / delete a table with inherited >> > children and RETURNING clause was added, is loses right >> > references to the pseudo columns, even though it works fine >> > without inherited children. >> > >> The attached patch fixed this known problem. > > This patch no longer applies to git master. Any chance of a rebase? > OK, I'll rebese it. > Also, might this approach work for the catalog? The use case I have > in mind is multi-tenancy, although one can imagine organizations where > internal access controls might be required on it, too. > If you intend to control behavior of DDL commands that internally takes access towards system catalog, RLS feature is not helpful. Please use sepgsql instead. :-) If you intend to control DML commands towards system catalogs, here is nothing special, so I expect it works as we are doing at user tables. Thanks, -- KaiGai Kohei <kaigai@kaigai.gr.jp>