Return-path: Received: from wf-out-1314.google.com ([209.85.200.169]:11717 "EHLO wf-out-1314.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755518AbYCEMaz (ORCPT ); Wed, 5 Mar 2008 07:30:55 -0500 Received: by wf-out-1314.google.com with SMTP id 28so1807105wff.4 for ; Wed, 05 Mar 2008 04:30:54 -0800 (PST) Message-ID: (sfid-20080305_123100_137036_DF78A3C6) Date: Wed, 5 Mar 2008 13:30:54 +0100 From: "Bart Van Assche" To: "Ingo Molnar" Subject: Re: [PATCH 6/9] drivers/net/wireless/iwlwifi/iwl-4965.c: Correct use of ! and & Cc: "Julia Lawall" , "Christopher Li" , yi.zhu@intel.com, linux-wireless@vger.kernel.org, ipw3945-devel@lists.sourceforge.net, linux-kernel@vger.kernel.org, kernel-janitors@vger.kernel.org, "Harvey Harrison" , "Alexander Viro" , linux-sparse@vger.kernel.org, "Josh Triplett" In-Reply-To: <20080305122010.GA999@elte.hu> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 References: <20080305063842.GA24495@elte.hu> <70318cbf0803042249j57d7f3a3j7666961a9132b10b@mail.gmail.com> <20080305070201.GA32434@elte.hu> <20080305122010.GA999@elte.hu> Sender: linux-wireless-owner@vger.kernel.org List-ID: On Wed, Mar 5, 2008 at 1:20 PM, Ingo Molnar wrote: > > * Julia Lawall wrote: > > > There are some legitimate uses of !x & y which are actually of the > > form !x & !y, where x and y are function calls. That is a not > > particularly elegant way of getting both x and y to be evaluated and > > then combining the results using "and". If such code is considered > > acceptable, then perhaps the sparse patch should be more complicated. > > i tend to be of the opinion that the details in C source code should be > visually obvious and should be heavily simplified down from what is > 'possible' language-wise - with most deviations and complications that > depart from convention considered an error. I'd consider "!fn1() & > !fn2()" a borderline coding style violation in any case - and it costs > nothing to change it to "!fn1() && !fn2()". If someone writes (!x & !y) instead of (!x && !y) because both x and y have to be evaluated, this means that both x and y have side effects. Please keep in mind that the C language does not specify whether x or y has to be evaluated first, so if x and y have to be evaluated in that order, an expression like (!x & !y) can be the cause of very subtle bugs. I prefer readability above brevity. Bart Van Assche.