From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 71977 invoked by alias); 2 May 2018 09:42:20 -0000 Mailing-List: contact gdb-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sourceware.org Received: (qmail 71054 invoked by uid 89); 2 May 2018 09:41:55 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-3.0 required=5.0 tests=AWL,BAYES_00,SPF_HELO_PASS autolearn=ham version=3.3.2 spammy=HCc:D*comcast.net, marchi, Marchi, HContent-Transfer-Encoding:8bit X-HELO: mx1.redhat.com Received: from mx3-rdu2.redhat.com (HELO mx1.redhat.com) (66.187.233.73) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Wed, 02 May 2018 09:41:52 +0000 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.rdu2.redhat.com [10.11.54.4]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 1C5068DC34; Wed, 2 May 2018 09:41:51 +0000 (UTC) Received: from [127.0.0.1] (ovpn04.gateway.prod.ext.ams2.redhat.com [10.39.146.4]) by smtp.corp.redhat.com (Postfix) with ESMTP id 5C7E72022C00; Wed, 2 May 2018 09:41:50 +0000 (UTC) Subject: Re: GDB 8.1 build error To: Simon Marchi References: <214C80CC-1173-41F6-AAA1-39C9D39E28B2@comcast.net> <454707570722fc0220074c0eca015a8f@polymtl.ca> <6171a043-e486-85ec-bdbb-2077a2b5ebd0@redhat.com> <9ed9617b-987e-3225-8518-a43eda0b5548@redhat.com> <13bcba4a8a70bcd977a7e644dd59e4bf@polymtl.ca> Cc: Paul Koning , gdb@sourceware.org From: Pedro Alves Message-ID: Date: Wed, 02 May 2018 09:42:00 -0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.6.0 MIME-Version: 1.0 In-Reply-To: <13bcba4a8a70bcd977a7e644dd59e4bf@polymtl.ca> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit X-SW-Source: 2018-05/txt/msg00002.txt.bz2 On 04/27/2018 09:41 PM, Simon Marchi wrote: > On 2018-04-27 16:24, Pedro Alves wrote: >> -Wnarrowing is on by default on gcc, and -w disables it. > > But it's a warning by default, not an error (like clang).  I think that's the important distinction. OK, That suggests to me that clang has -Werror=narrowing enabled by default, instead of -Wnarrowing. But I still think that the real issue is that gcc and clang behave differently wrt "-w" precedence. Note, with: $ cat narrow.cc int return_int () { return 42; } char buf[2] = { return_int () }; #1 - even if you make gcc error out with -Werror=narrowing, a subsequent "-w" cancels the error/warning, not so with clang: $ g++ -std=gnu++17 narrow.cc -c -Werror=narrowing -w $ clang++ -std=gnu++17 narrow.cc -c -Werror=narrowing -w narrow.cc:2:17: error: non-constant-expression cannot be narrowed from type 'int' to 'char' in initializer list [-Wc++11-narrowing] #2 - If you put "-Wno-error=narrowing", after the "-w", then both compilers suppress the warning/error: $ clang++ -std=gnu++17 narrow.cc -c -Werror=narrowing -w -Wno-error=narrowing $ g++ -std=gnu++17 narrow.cc -c -Werror=narrowing -w -Wno-error=narrowing #3 - For completeness, adding "-Werror=narrowing" after the -w, shows #1 again: $ g++ -std=gnu++17 narrow.cc -c -w -Werror=narrowing $ clang++ -std=gnu++17 narrow.cc -c -w -Werror=narrowing narrow.cc:2:17: error: non-constant-expression cannot be narrowed from type 'int' to 'char' in initializer list [-Wc++11-narrowing] $ g++ -std=gnu++17 narrow.cc -c -Werror=narrowing -w -Werror=narrowing $ clang++ -std=gnu++17 narrow.cc -c -Werror=narrowing -w -Werror=narrowing narrow.cc:2:17: error: non-constant-expression cannot be narrowed from type 'int' to 'char' in initializer list [-Wc++11-narrowing] So it seems to be the issue here is more about "-w" precedence over all warning switches than about what is enabled by default. To confirm this, we can see the same thing with any other warning: ~~~~~~~~~~~~ $ gcc unused.c -c -Werror=unused-variable unused.c: In function ‘function’: unused.c:1:23: error: unused variable ‘i’ [-Werror=unused-variable] int function () { int i; return 42; } ^ cc1: some warnings being treated as errors ~~~~~~~~~~~~ ~~~~~~~~~~~~ $ clang unused.c -c -Werror=unused-variable unused.c:1:23: error: unused variable 'i' [-Werror,-Wunused-variable] int function () { int i; return 42; } ^ 1 error generated. ~~~~~~~~~~~~ vs $ gcc unused.c -c -Werror=unused-variable -w $ clang unused.c -c -Werror=unused-variable -w unused.c:1:23: error: unused variable 'i' [-Werror,-Wunused-variable] So seems like we could handle this my making --disable-build-warnings use "-Wno-error -w" instead of just "-w". But I'd suggest checking with clang and/or gcc folks to confirm the difference is intentional too. Thanks, Pedro Alves