From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 1825 invoked by alias); 6 Mar 2003 08:53:49 -0000 Mailing-List: contact rhdb-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Post: List-Help: , Sender: rhdb-owner@sources.redhat.com Received: (qmail 1817 invoked from network); 6 Mar 2003 08:53:47 -0000 Received: from unknown (HELO mail.gizmolab.net) (63.193.253.82) by 172.16.49.205 with SMTP; 6 Mar 2003 08:53:47 -0000 Received: from gizmolab.com (woozy.gizmolab.net [10.1.1.3]) by mail.gizmolab.net (Postfix) with ESMTP id E9FEE479A5 for ; Thu, 6 Mar 2003 01:43:54 -0800 (PST) Date: Thu, 06 Mar 2003 08:53:00 -0000 Mime-Version: 1.0 (Apple Message framework v551) Content-Type: text/plain; charset=US-ASCII; format=flowed Subject: postgresql 7.3 From: Jonathan Sand To: rhdb@sources.redhat.com Content-Transfer-Encoding: 7bit Message-Id: X-SW-Source: 2003-q1/txt/msg00048.txt.bz2 can you give me an approximate timeframe for redhat's release of rhdb based on postgresql 7.3? my problem is that I'm using redhat bugzilla 2.15 with a database created with bugzilla based on mysql. I've managed to transfer the data out of mysql and into postgres (no small feat), and bugzilla runs, but it has a few bugs. I've, since then, downloaded redhat bugzilla 2.17, only to discover that the schema expected in this version are quite different. among the many changes, one of them can't be performed by rhdb 2.1 (postgresql 7.2.3). specifically: several of the tables contain essentially identical data (eg, products.value and bugs.product) with data type varchar. this being inefficient as a primary key and foreign key, these column definitions are being replaced by integer primary keys and integer references, respectively. postgres 7.3 allows the following sql: alter table products add id int; ... set new unique values in this new id column .... alter table products alter id set not null; alter table products add primary key(id); postgresql 7.2.3 doesn't allow that the "set not null" column constraint be applied in an "alter" statement and the "add primary key" can only be applied to columns which are designated as "not null." a neat and frustrating shutout. I suppose I could dump the database and reconstruct the schema correctly, but I'd then have to seriously modify the redhat bugzilla setup code. moral of this story? the bleeding edge is not for whimps. sand@gizmolab.com