From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 29712 invoked by alias); 5 Feb 2013 16:47:01 -0000 Received: (qmail 29619 invoked by uid 22791); 5 Feb 2013 16:46:59 -0000 X-SWARE-Spam-Status: No, hits=-3.8 required=5.0 tests=AWL,BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,KHOP_RCVD_TRUST,KHOP_THREADED,RCVD_IN_DNSWL_LOW,RCVD_IN_HOSTKARMA_YE X-Spam-Check-By: sourceware.org Received: from mail-qc0-f171.google.com (HELO mail-qc0-f171.google.com) (209.85.216.171) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 05 Feb 2013 16:46:55 +0000 Received: by mail-qc0-f171.google.com with SMTP id d1so140207qca.30 for ; Tue, 05 Feb 2013 08:46:53 -0800 (PST) MIME-Version: 1.0 X-Received: by 10.49.73.232 with SMTP id o8mr23736512qev.0.1360082814039; Tue, 05 Feb 2013 08:46:54 -0800 (PST) Received: by 10.49.131.102 with HTTP; Tue, 5 Feb 2013 08:46:53 -0800 (PST) In-Reply-To: <20130205093846.GH5023@bubble.grove.modra.org> References: <1356420600-11507-1-git-send-email-vapier@gentoo.org> <1358483201-12733-1-git-send-email-vapier@gentoo.org> <1358483201-12733-2-git-send-email-vapier@gentoo.org> <20130205093846.GH5023@bubble.grove.modra.org> Date: Tue, 05 Feb 2013 16:47:00 -0000 Message-ID: Subject: Re: [PATCH 2/2 v5] gold: enable new dtags by default From: "H.J. Lu" To: Ian Lance Taylor , "H.J. Lu" , Mike Frysinger , binutils@sourceware.org Content-Type: text/plain; charset=ISO-8859-1 X-IsSubscribed: yes Mailing-List: contact binutils-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: binutils-owner@sourceware.org X-SW-Source: 2013-02/txt/msg00041.txt.bz2 On Tue, Feb 5, 2013 at 1:38 AM, Alan Modra wrote: > On Mon, Feb 04, 2013 at 09:43:04PM -0800, Ian Lance Taylor wrote: >> It seems that we have made sensible-seeming decisions to wind up in an >> absurd place. > > Right, so where did we go wrong? I think it was in assuming that we > could default to --enable-new-dtags. Users were affected. (I'm > assuming HJ's PR ld/15096 came from a real world problem.) > > I propose reverting HJ's patch, and Mike's and Roland's 2013-02-21 > patches. That will leave us with just Mike's 2013-01-18 change to > disable old dtags when new dtags are selected. If people want new > dtags by default, do so via the gcc driver. > Or we can just revert my change and add --runpath. -rpath will always generate DT_RPATH and --runpath will generate DT_RUNPATH. What should happen when both -rpath and --runpath are on command line? -- H.J.