From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Korsgaard Date: Sun, 11 Dec 2011 23:41:44 +0100 Subject: [Buildroot] [PATCH 3/3] libglib2: bump to version 2.30.2 In-Reply-To: <1323171959-24873-4-git-send-email-zonque@gmail.com> (Daniel Mack's message of "Tue, 6 Dec 2011 12:45:59 +0100") References: <1323171959-24873-1-git-send-email-zonque@gmail.com> <1323171959-24873-4-git-send-email-zonque@gmail.com> Message-ID: <874nx6bujb.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 >>>>> "Daniel" == Daniel Mack writes: Daniel> Important changes for this version bump are that libffi (for both host Daniel> and target) is now a hard requirement for libglib. Daniel> The patch regarding atomic function seems obsolete now, as this part of Daniel> the lib was rewritten. I couldn't figure which platform this used to Daniel> break on, but in worst case, it would be trivial to redo this one-liner. Daniel> Also, there's an issue in the upstream package that makes configure fail Daniel> for cross-builds with "configure: error: cannot run test program while Daniel> cross compiling". This issue was reported, but the patch was rejected. Daniel> Appearantly, the developers don't expect people to use configure for Daniel> cross environments anymore, so we have to maintain this patch locally. Daniel> For more information, see Thanks. Instead of using this patch I would prefer to just pass glib_cv_have_qsort_r=yes/no in LIBGLIB2_CONF_ENV like Colin suggested. That way we can set it to yes for external glibc-based toolchains (BR2_TOOLCHAIN_EXTERNAL_GLIBC) and no otherwise. Care to respin this patch with that change? Thanks. -- Bye, Peter Korsgaard