Hi developers!
Testing the pgAdmin III 1.8.0 Beta 3 (Aug 10 2007, rev: 6546M). Client
Win XP, host: Debian Etch / PG 8.2.4 and Debian Sarge / PG 8.1.8
The reverse engineered SQL for superusers seems at fault. Applies to
both pg 8.2 and pg 8.1.
First example: everything looks normal:
CREATE ROLE tester;
=== quote
-- Role: "tester"
-- DROP ROLE tester;
CREATE ROLE tester
NOSUPERUSER INHERIT NOCREATEDB NOCREATEROLE;
=== unquote
Second example: superusers get an extra line "UPDATE ..." that should
not be there. Especially as it refers to the OID which would be
different when recreating the user.
CREATE ROLE tester2 SUPERUSER;
=== quote
-- Role: "tester2"
-- DROP ROLE tester2;
CREATE ROLE tester2
SUPERUSER INHERIT NOCREATEDB NOCREATEROLE;
UPDATE pg_authid SET rolcatupdate=true WHERE OID=3714960::oid;
=== unquote
Regards
Erwin
---------------------------(end of broadcast)---------------------------
TIP 4: Have you searched our list archives?
http://archives.postgresql.org
