From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Korsgaard Date: Mon, 12 Dec 2011 20:43:20 +0100 Subject: [Buildroot] Make problem on 64-bit linux with CS 2009q1 In-Reply-To: (Trevor Woerner's message of "Mon, 12 Dec 2011 12:52:15 -0500") References: <4EE1CB14.9010802@free.fr> <20111209193729.0cc88e6e@skate> <87y5ull8z0.fsf@macbook.be.48ers.dk> <87ty59l5y1.fsf@macbook.be.48ers.dk> <4EE5B9E9.40404@free.fr> <87hb16up5z.fsf@macbook.be.48ers.dk> <4EE5D3BB.7000700@free.fr> <20111212111750.0473fb52@skate> Message-ID: <87d3bttw2v.fsf@macbook.be.48ers.dk> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net >>>>> "Trevor" == Trevor Woerner writes: Hi, Trevor> It appears that the contents of openSUSE 12.1's site file are buggy. Trevor> The CONFIG_SITE mechanism is purposefully provided to help "prime the Trevor> cache" for ./configure scripts when cross compiling to provide answers Trevor> to questions which would normally require code to be compiled and run Trevor> (which fails, of course, when the compiler being used is a cross Trevor> compiler). Trevor> Should the CONFIG_SITE feature be disabled and otherwise unavailable Trevor> to users just because there's a bug in openSUSE's implementation? Yes, because whatever defaults potentially could make sense for the build host in general doesn't make sense for the target, and for BR builds to reproducable behaviour shouldn't change depending on something outside our control. You could perhaps argue for allowing it for host packages, but the simplest/safest option is just to disallow it for all. -- Bye, Peter Korsgaard