From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263085AbUG0KJ6 (ORCPT ); Tue, 27 Jul 2004 06:09:58 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263117AbUG0KJ4 (ORCPT ); Tue, 27 Jul 2004 06:09:56 -0400 Received: from mail.gmx.de ([213.165.64.20]:19106 "HELO mail.gmx.net") by vger.kernel.org with SMTP id S261951AbUG0KJw (ORCPT ); Tue, 27 Jul 2004 06:09:52 -0400 X-Authenticated: #1725425 Date: Tue, 27 Jul 2004 12:10:40 +0200 From: Marc Ballarin To: Junio C Hamano Cc: mpm@selenic.com, linux-kernel@vger.kernel.org Subject: Re: [PATCH] Delete cryptoloop Message-Id: <20040727121040.27ca7170.Ballarin.Marc@gmx.de> In-Reply-To: <7v4qntc06o.fsf@assigned-by-dhcp.cox.net> References: <20040726200126.GQ5414@waste.org> <7v4qntc06o.fsf@assigned-by-dhcp.cox.net> X-Mailer: Sylpheed version 0.9.12 (GTK+ 1.2.10; i686-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Tue, 27 Jul 2004 01:40:47 -0700 Junio C Hamano wrote: > > Jari's exploit uses the property that his watermarks are > encrypted to identical ciphertext blocks, but does it mean that > the technique can be used to prove that identical ciphertext are > from the watermarks and not coming from mere coincidence? > You can't prove anything by a single found watermark, but you can encode additional information by placing the watermarks at chosen offsets in your file. For example, take prime numbers starting at 2 as offsets, and place the encodings in a chosen order at those offsets. The chances for a random occurence of this pattern are marginal, and - unless the watermarked file becomes strongly fragmented - this pattern is easy to retrieve. While this might not be evidence in the legal sense (it could well be, but IANAL), it is at the very least a strong indication. Regards