From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hamish Moffatt Date: Tue, 16 Dec 2008 11:37:56 +1100 Subject: [Buildroot] svn commit: trunk/buildroot/package In-Reply-To: <20081215234234.575aca77@surf> References: <20081215220738.6F7173C6BA@busybox.net> <20081215234234.575aca77@surf> Message-ID: <20081216003756.GA19127@cloud.net.au> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net On Mon, Dec 15, 2008 at 11:42:34PM +0100, Thomas Petazzoni wrote: > Le Mon, 15 Dec 2008 14:07:38 -0800 (PST), > tpetazzoni at uclibc.org a ?crit : > > > TARGET_PATH didn't contain $(STAGING_DIR)/usr/bin, which means that > > programs installed in $(STAGING_DIR)/usr/bin were not considered for > > execution during Buildroot build process. This was a problem with host > > automake/autoconf/libtool, which could not be found. > > BTW, I'm quite confused by the fact that we install host binaries and > target binaries in $(STAGING_DIR). It means that during the build, we > have target binaries considered for execution because they are in > $(STAGING_DIR)/usr/bin, which is the the PATH during Buildroot > execution. > > Is there any reason for this ? > > I would really find it cleaner to have a separate installation > directory for host tools (like toolchain_build_ARCH/bin which contains > a symlink to sed). Yes it's a mess. There's host binaries like the compiler etc, and some target binaries provided by various library packages eg openssl, libjpeg.. We need to clean up the Python build situation too (eg build a proper host Python for use in building target Python), but installing all of that into the existing $(STAGING_DIR) will just make it worse. Hamish -- Hamish Moffatt VK3SB