From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 5142 invoked by alias); 8 Sep 2014 15:37:25 -0000 Mailing-List: contact libc-alpha-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: libc-alpha-owner@sourceware.org Received: (qmail 5122 invoked by uid 89); 8 Sep 2014 15:37:22 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-3.9 required=5.0 tests=AWL,BAYES_00,RP_MATCHES_RCVD,SPF_HELO_PASS autolearn=ham version=3.3.2 X-HELO: mx1.redhat.com Message-ID: <540DCD2E.7060107@redhat.com> Date: Mon, 08 Sep 2014 15:37:00 -0000 From: "Carlos O'Donell" User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.7.0 MIME-Version: 1.0 To: "Joseph S. Myers" CC: GNU C Library Subject: Re: In debug/tst-chk1 we purposely test gets and getwd? References: <540D0803.5030502@redhat.com> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit X-SW-Source: 2014-09/txt/msg00111.txt.bz2 On 09/08/2014 11:28 AM, Joseph S. Myers wrote: > On Sun, 7 Sep 2014, Carlos O'Donell wrote: > >> It seems in debug/tst-chk1.c we purposely test gets >> and getwd, but both of those trigger deprecation >> warnings. It would be nice to eventually turn on >> -Werror. > > Yes, for various testcases it will be necessary to disable some warnings, > or to use -Wno-error= for them, because tests need to cover code using > deprecated interfaces, and code doing bad things that can be detected at > compile time (e.g. some _FORTIFY_SOURCE tests). > > There are already various -Wno- settings in the makefiles which it may > make sense to review to see if they are still necessary or if there are > better ways to address those warnings (if still present) - we'll need to > work out our policy on when to use such settings, and how to handle > hard-to-fix warnings in general, when using -Werror by default. > > (At one point I thought maybe we should enable -Werror at first only for > installed code, not tests. But I now think it would be simpler to enable > it everywhere and then selectively disable it for particular tests; > probably most warnings in tests are in fact easy to fix without affecting > what's being tested.) Thanks for that feedback. Any particular opposition to #pragma usage? It seems like easier maintenance to add the #pragma's close to their point of use with comments talking about why we avoid the warning. Cheers, Carlos.