From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Korsgaard Date: Fri, 25 Apr 2008 10:21:38 +0200 Subject: [Buildroot] svn commit: trunk/buildroot/package: oprofile In-Reply-To: <20080425064853.GA1525@cloud.net.au> (Hamish Moffatt's message of "Fri\, 25 Apr 2008 16\:48\:53 +1000") References: <20080424165429.5E3993C7F0@busybox.net> <87zlrj2khz.fsf@macbook.be.48ers.dk> <20080425064853.GA1525@cloud.net.au> Message-ID: <87fxtaxs31.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 >>>>> "Hamish" == Hamish Moffatt writes: Hi, >> Please use the .patch.avr32 naming convention. Hamish> Isn't that only required if the patch is arch-specific, ie it Hamish> breaks other architectures? A quick look at the patch Hamish> suggests to me that it adds avr32 support to profile without Hamish> breaking anything else. Ie it's safe to always apply it. Maybe, I didn't look at it. Anyway, the safest solution is use the .patch.avr32 convention, then it certainly cannot harm other archs. >> Does that actually work? I mean, does the compiler get recompiled with >> C++ support if you select this package after your initial make? Hamish> I doubt it.. I think it's quite difficult to get gcc recompiled after Hamish> changing the options - usually means deleting most (if not all) of Hamish> toolchain_build_$arch. Then it should be a depends instead. Anyway, a select here isn't that nice from a usability POV anyway. I would prefer to get a warning before buildroot would start compiling in C++ support from me enabling a random package. -- Bye, Peter Korsgaard