From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Sun, 19 Feb 2017 18:44:32 +0100 Subject: [Buildroot] [PATCH 4/4] bctoolbox: disable rpath In-Reply-To: <1487516855.17793.1.camel@embedded.rocks> References: <20170214224402.8785-1-joerg.krause@embedded.rocks> <20170214224402.8785-4-joerg.krause@embedded.rocks> <20170215052846.yxjzvhmsgdc3hy57@tarshish> <20170215095650.236f21b5@free-electrons.com> <1487516855.17793.1.camel@embedded.rocks> Message-ID: <20170219184432.58c82ea1@free-electrons.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Hello, On Sun, 19 Feb 2017 16:07:35 +0100, J?rg Krause wrote: > > Worth mentioning that there is both CMAKE_SKIP_RPATH and > > CMAKE_SKIP_INSTALL_RPATH. The former skips adding the rpath during > > both > > the build and install steps, which prevents from running the binary > > from the build directory. The latter only skips the rpath during the > > install step. But since we're cross-compiling, running on the host > > machine from the build directory anyway doesn't make sense. So > > CMAKE_SKIP_RPATH is OK. > > > > See https://cmake.org/cmake/help/v3.0/variable/CMAKE_SKIP_RPATH.html. > > Samuel argues not to force this option by the infrastructure. I am > not sure how do we handle this properly. So, what shall we do? As I said in my second e-mail (sent after reading Samuel's answer), Samuel is probably correct. Skipping the rpath in *all* cases may not be correct, so we have to do it at the per-package level, like your patch does. Thomas -- Thomas Petazzoni, CTO, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com