From owner-freebsd-hackers@FreeBSD.ORG Tue Feb 10 08:11:27 2004 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 27A0C16A4CE for ; Tue, 10 Feb 2004 08:11:27 -0800 (PST) Received: from kientzle.com (h-66-166-149-50.SNVACAID.covad.net [66.166.149.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id E260043D2F for ; Tue, 10 Feb 2004 08:11:26 -0800 (PST) (envelope-from kientzle@acm.org) Received: from acm.org ([66.166.149.54]) by kientzle.com (8.12.9/8.12.9) with ESMTP id i1AGBLkX013437; Tue, 10 Feb 2004 08:11:24 -0800 (PST) (envelope-from kientzle@acm.org) Message-ID: <402902A8.8040005@acm.org> Date: Tue, 10 Feb 2004 08:11:20 -0800 From: Tim Kientzle User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.4) Gecko/20031006 X-Accept-Language: en-us, en MIME-Version: 1.0 To: =?ISO-8859-1?Q?Dag-Erling_Sm=F8rgrav?= References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit cc: hackers@freebsd.org Subject: Re: how to fool gcc? X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: kientzle@acm.org List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 10 Feb 2004 16:11:27 -0000 Suggestions: 1) Convert RETURNS(s) into a function rather than a macro. 2) Use a ternary (someone else suggested this) 3) Remove the printf-like declaration (Yes, this is distasteful.) 4) Add an intermediate variable: do { const char *_t = (s) == NULL ? "NULL" : (s); openpam_log(... , _t); return (s); } while (0) That's all I can think of right now, though I'm certain there are numerous other fixes. Good luck, (let us know what works!) Tim Dag-Erling Smørgrav wrote: > I'm having trouble with some uncommitted OpenPAM patches that I'd like > to get into the tree. The problem actually doesn't occur with a > normal build, but it prevents me from building a debugging version of > libpam. > > Part of the patch declares openpam_log(3) as printf-like so gcc can > check format strings etc. However, openpam_log(3) is also used in > debugging macros such as this: > > #define RETURNS(s) do { \ > if ((s) == NULL) \ > openpam_log(PAM_LOG_DEBUG, "returning NULL"); \ > else \ > openpam_log(PAM_LOG_DEBUG, "returning '%s'", (s)); \ > return (s); \ > } while (0) > > The problem is that when it encounters RETURNS(NULL), gcc complains > that I'm passing a NULL argument to printf(3), even though it should > be obvious that I'm not: > > cc -O -pipe -march=pentium2 -I/usr/src/lib/libpam/libpam -I/home/des/projects/openpam/include -DLIB_MAJ=2 -g -DDEBUG -Wsystem-headers -Werror -Wall -Wno-format-y2k -W -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Wreturn-type -Wcast-qual -Wwrite-strings -Wswitch -Wshadow -Wcast-align -Wbad-function-cast -Wchar-subscripts -Winline -Wnested-externs -Wredundant-decls -c /home/des/projects/openpam/lib/openpam_get_option.c > /home/des/projects/openpam/lib/openpam_get_option.c: In function `openpam_get_option': > /home/des/projects/openpam/lib/openpam_get_option.c:62: warning: reading through null pointer (arg 4) > /home/des/projects/openpam/lib/openpam_get_option.c:73: warning: reading through null pointer (arg 4) > *** Error code 1 > > Stop in /usr/src/lib/libpam/libpam. > > I've tried various twists to fool gcc, such as casting (s) to (const > char *) and adding 0 to it hoping that the addition would defeat its > NULL pointer check. Nothing I've tried works, though, and I would > really hate to have to lower the WANRS level just for this. > > Any suggestions? > > DES