From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 6685 invoked by alias); 7 May 2003 18:06:12 -0000 Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-owner@gcc.gnu.org Received: (qmail 6630 invoked from network); 7 May 2003 18:06:09 -0000 Received: from unknown (HELO piper.synopsys.com) (204.176.21.196) by sources.redhat.com with SMTP; 7 May 2003 18:06:09 -0000 Received: (from jbuck@localhost) by piper.synopsys.com (8.11.6/8.11.6) id h47I61E15177; Wed, 7 May 2003 11:06:01 -0700 Date: Wed, 07 May 2003 18:06:00 -0000 From: Joe Buck To: Benjamin Kosnik Cc: gcc@gcc.gnu.org Subject: Re: draft release notes for gcc 3.3 Message-ID: <20030507110601.A8072@synopsys.com> References: <20030507124939.0da846eb.bkoz@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: <20030507124939.0da846eb.bkoz@redhat.com>; from bkoz@redhat.com on Wed, May 07, 2003 at 12:49:39PM -0500 X-SW-Source: 2003-05/txt/msg00667.txt.bz2 On Wed, May 07, 2003 at 12:49:39PM -0500, Benjamin Kosnik wrote: > > (Need to say something about C++ binary compatibility, versioning of > > libstdc++ etc). > > It's my thought that you should say that 3.3.x is C++ ABI compatible > with 3.2.x. I believe this is certainly the intention. OK, let's make sure that I can explain all the issues correctly. As far as the compiler is concerned, we have ABI compatibility. For the library, we've made many changes, but use versioning so that programs linked against 3.2.x's libstdc++ will continue to work. But this only works on OSes that implement versioning, correct? Folks on at least some OSes will need to avoid mixing. > > 8463 std::ios_base has a non-virtual destructor > > 8463 is not fixed, nor fixable, for 3.3 due to ABI constraints, sadly. > It is fixed on mainline. Everything else looks ok though. Must have misread something; I will fix this.