From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 2988 invoked by alias); 2 May 2005 09:05:33 -0000 Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org Received: (qmail 2955 invoked by uid 48); 2 May 2005 09:05:31 -0000 Date: Mon, 02 May 2005 09:05:00 -0000 Message-ID: <20050502090531.2954.qmail@sourceware.org> From: "ebotcazou at gcc dot gnu dot org" To: gcc-bugs@gcc.gnu.org In-Reply-To: <20050429073316.21277.info@yourkit.com> References: <20050429073316.21277.info@yourkit.com> Reply-To: gcc-bugzilla@gcc.gnu.org Subject: [Bug target/21277] Runtime error with C++ shared library and --disable-shared X-Bugzilla-Reason: CC X-SW-Source: 2005-05/txt/msg00218.txt.bz2 List-Id: ------- Additional Comments From ebotcazou at gcc dot gnu dot org 2005-05-02 09:05 ------- > We must link our .so statically with all the gcc stuff to make sure it runs on > every Solaris. Shipping libstd++.so with our shared library is not very > suitable for us, because it makes product download size bigger. 5 MB uncompressed for 32-bit, 6 MB uncompressed for 64-bit! > Anyway, option --disable-shared exists, and is documented. So it should either > properly work (for platforms it is supported for), or be dropped (forplatforms > where it is not supported). While there's nothing said it is not supported for > Solaris, all its improper work is a bug, and nothing but a bug. Isn't it? :) There is nothing wrong in the current behavior of --disable-shared: it builds static libraries the way static libraries should be built. Your practice of building shared libraries with a compiler configured with --disable-shared looks far more questionable to me; if I were to change something, I'd simply reject -shared in that case. Note for example that a shared libgcc is required on Solaris for exception propagation accross shared libraries. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21277