From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from cantor2.suse.de ([195.135.220.15]:36659 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753674AbbGBUG3 (ORCPT ); Thu, 2 Jul 2015 16:06:29 -0400 Message-ID: <559599C2.6090505@suse.cz> Date: Thu, 02 Jul 2015 22:06:26 +0200 From: Michal Marek MIME-Version: 1.0 Subject: Re: [PATCH 0/2] kconfig: allow use of relations other than (in)equality References: <557EDA310200007800084C43@mail.emea.novell.com> <6743206.QpmKW5snMP@tacticalops> In-Reply-To: <6743206.QpmKW5snMP@tacticalops> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Sender: linux-kbuild-owner@vger.kernel.org List-ID: To: Martin Walch Cc: Jan Beulich , pebolle@tiscali.nl, linux-kbuild@vger.kernel.org, linux-kernel@vger.kernel.org Dne 2.7.2015 v 15:04 Martin Walch napsal(a): > Hi, > >> 1: allow use of relations other than (in)equality > > I know it is a bit late for objections. Still, I want to point out that > this looks to me like a major extension to the language. > > Kconfig is a configuration language, and as far as I can tell it is > (intentionally) not Turing complete to keep the configuration simple and > controllable. All relations that have been defined so far check for equality > (or for being not equal). The new relations "<=", ">=", "<", and ">" add more > expressiveness, potentially making the language actually more complex and > reasoning harder. The patch just adds four new binary operations of the same order as the existing == and !=, with a the semantics that everybody expects. And the grammar for kconfig expressions is so simplistic that you cannot even write things like (A && B) == (C && D). So turing completeness is not a topic here, neither before nor after this patch. Michal