From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnout Vandecappelle Date: Wed, 15 Jul 2015 14:43:03 +0200 Subject: [Buildroot] [PATCH 1/1] toolchain: create symlink to 'lib' from ARCH_LIB_DIR iso fixed lib32/lib64 In-Reply-To: References: <1436949545-12487-1-git-send-email-patrickdepinguin@gmail.com> <20150715105526.19c500e8@free-electrons.com> <20150715105950.42ac4b53@free-electrons.com> <55A62262.1090304@mind.be> Message-ID: <55A65557.5030709@mind.be> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net On 07/15/15 12:08, Thomas De Schampheleire wrote: > On Wed, Jul 15, 2015 at 11:05 AM, Arnout Vandecappelle wrote: >> On 07/15/15 10:59, Thomas Petazzoni wrote: >>> Hello, >>> >>> On Wed, 15 Jul 2015 10:57:52 +0200, Thomas De Schampheleire wrote: >>> >>>> Well that is part of the discussion: conceptually I believe that the >>>> symlink creation is part of the toolchain as a whole. It needs to be >>>> done before libc, but actually all remaining toolchain components will >>>> probably rely on it, and also the packages. >>>> By placing it in gcc-initial I believe it is pretty hidden. >>> >>> It is also pretty hidden to put the hook for a given package outside of >>> this package .mk file. >>> >>>> Moreover, suppose we add support for the llvm/clang compiler. For such >>>> toolchains, we'd also need the symbolic links, so that code would have >>>> to be duplicated... >>> >>> I don't think trying to take llvm/clang into account really makes sense >>> here. It's probably going to be so much different than what we have >>> today that lots of things would have to be reworked anyway. >> >> Just to be clear: for me this is a bikeshedding discussion. Neither is right, >> so just go for one. >> >> BTW, yesterday I did ask everyone in Toulouse if it would be OK have add to the >> gcc-final hook from toolchain-buildroot.mk, and nobody really opposed (though I >> don't think you really said anything, ThomasP). >> > > Suppose we do add the hook in gcc-initial, where would the link target > be defined? So this code: > > +# MIPS64/n32 requires lib32 even though it's a 64-bit arch. > +ifeq ($(BR2_ARCH_IS_64)$(BR2_MIPS_NABI32),y) > +TOOLCHAIN_BUILDROOT_LIB_SYMLINK = lib64 > +else > +TOOLCHAIN_BUILDROOT_LIB_SYMLINK = lib32 > +endif > > > Do you think that belongs also in gcc-initial? Or should that stay in > buildroot-toolchain? IMHO it should certainly stay together with the piece of code that actually uses it - which is the point of putting it into gcc-initial to begin with (otherwise the adding of the hook is not in the same file as the rest of gcc-initial). And of course, it should be called GCC_INITIAL_* then. Regards, Arnout -- Arnout Vandecappelle arnout dot vandecappelle at essensium dot com Senior Embedded Software Architect . . . . . . +32-478-010353 (mobile) Essensium, Mind division . . . . . . . . . . . . . . http://www.mind.be G.Geenslaan 9, 3001 Leuven, Belgium . . . . . BE 872 984 063 RPR Leuven LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle GPG fingerprint: 7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF