From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [85.37.17.114] (helo=smtp-out114.alice.it) by linuxtogo.org with esmtp (Exim 4.69) (envelope-from ) id 1MKjeX-0002Dm-AJ for openembedded-devel@lists.openembedded.org; Sun, 28 Jun 2009 03:57:57 +0200 Received: from FBCMMO02.fbc.local ([192.168.68.196]) by smtp-out114.alice.it with Microsoft SMTPSVC(6.0.3790.3959); Sun, 28 Jun 2009 03:46:22 +0200 Received: from FBCMCL01B04.fbc.local ([192.168.69.85]) by FBCMMO02.fbc.local with Microsoft SMTPSVC(6.0.3790.1830); Sun, 28 Jun 2009 03:46:22 +0200 Received: from [192.168.1.171] ([79.51.163.91]) by FBCMCL01B04.fbc.local with Microsoft SMTPSVC(6.0.3790.3959); Sun, 28 Jun 2009 03:46:22 +0200 From: GNUtoo To: openembedded-devel@lists.openembedded.org In-Reply-To: <6a1b49480906251521w364bce23o2dd5f921a468474c@mail.gmail.com> References: <6a1b49480906251521w364bce23o2dd5f921a468474c@mail.gmail.com> Date: Sun, 28 Jun 2009 03:46:13 +0200 Message-Id: <1246153579.32326.3.camel@port4> Mime-Version: 1.0 X-Mailer: Evolution 2.24.4 X-OriginalArrivalTime: 28 Jun 2009 01:46:22.0515 (UTC) FILETIME=[3D754030:01C9F792] Subject: Re: patch to boost package X-BeenThere: openembedded-devel@lists.openembedded.org X-Mailman-Version: 2.1.11 Precedence: list Reply-To: openembedded-devel@lists.openembedded.org List-Id: Using the OpenEmbedded metadata to build Distributions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 28 Jun 2009 01:57:57 -0000 Content-Type: text/plain Content-Transfer-Encoding: 7bit On Fri, 2009-06-26 at 00:21 +0200, Frank Black wrote: > I tried building boost 1.36.0 for my armv5te processor. Problem is there is > no support for the function "__sync_add_and_fetch" called from > detail/atomic_count_sync.hpp. There are two patches which both fails to > resolve this for me. > > 1: arm-instrinct.patch. This requires __ARM_ARCH_7A__ to be defined, which > is not good for me on an armv5te. > 2: atomic_count_gcc_atomicity.patch: Works, however problem is that > atomic_count.hpp contains #defines that always #includes > atomic_count_sync.hpp instead of atomic_count_gcc.hpp, so the patch has no > effect > > I solved this by applying a really ugly patch (see below) that makes sure > atomic_count_sync.hpp is never used. Im sure therre is a better way to do it > but I'm not at all familiar what kind of ARM-specific defines you got to > play around with. Maybe someone has more in-depth knowledge abot this? > > > > --- boost_1_33_1.orig/boost/detail/atomic_count.hpp 2007-11-23 > 18:03:14.000000000 +0100 > +++ boost_1_33_1/boost/detail/atomic_count.hpp 2009-06-25 > 23:07:10.000000000 +0200 > @@ -101,10 +101,10 @@ > > # include > > -#elif defined( __GNUC__ ) && ( __GNUC__ * 100 + __GNUC_MINOR__ >= 401 ) > +/*#elif defined( __GNUC__ ) && ( __GNUC__ * 100 + __GNUC_MINOR__ >= 401 ) > > # include > - > +*/ > #elif defined(__GLIBCPP__) || defined(__GLIBCXX__) > > # include > > > / Rolle yes there is: I've put the boost work I did with the help of other people here: http://gnutoo.homelinux.org/open_embedded/boost.tar.bz2 I didn't have the time to finish my work because I had a segmentation fault in boost with wesnoth-1.6x so I didn't commit it yet because of the lack of skills and time to debug it(I don't know C++,I just learned it a bit for this precise problem) Denis.