Re: patch with no testing

From: Paul Drake <drak0nian_at_yahoo.com>
Date: 6 May 2003 22:29:36 -0700
Message-ID: <1ac7c7b3.0305062129.4b379810@posting.google.com>

"Terry Dykstra" <dontreply_tddykstra_at_forestoil.ca> wrote in message news:<gkvsa.21094$yv1.1564678_at_news2.telusplanet.net>...
> 8.1.7.4.9 has the fixes for security alerts 48-52 and it is already
> superseded by 8.1.7.4.10.
>
> --
> Terry Dykstra
> Canadian Forest Oil Ltd.

Terry,

IF (and not only IF) you are using cursor_sharing = FORCE ... you could be in for a nasty surprise after applying 8.1.7.4.10.

This is exactly why you have to test patchsets prior to putting them into production. One off patchsets are not regression tested by Oracle, regardless of if there are seemingly incredibly important security fixes that one might think could force a major patchset release. It won't.

The patchset that I'm now testing is 9.2.0.3.

done soon with 8.1.7,

Paul Received on Wed May 07 2003 - 00:29:36 CDT