From mboxrd@z Thu Jan 1 00:00:00 1970 From: Luca Ceresoli Date: Mon, 18 Feb 2013 20:37:15 +0100 Subject: [Buildroot] [autobuild.buildroot.net] Build results for 2013-02-12 In-Reply-To: <20130215055400.6688c117@skate> References: <20130213073425.9C3C052C6D2@lolut.humanoidz.org> <511D556D.3060503@lucaceresoli.net> <20130215055400.6688c117@skate> Message-ID: <512282EB.7070702@lucaceresoli.net> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Thomas Petazzoni wrote: > Dear Luca Ceresoli, > > On Thu, 14 Feb 2013 22:21:49 +0100, Luca Ceresoli wrote: > >> Should I send a patch to set -no-pch? I'm reluctant because it's not a real >> fix, and compile time could increase significantly, I guess. >> >> Thomas, this probably affects Qt5 Beta 1 as well, according to this blog >> post: >> http://gpupowered.org/node/14. Would you mind testing it on your Qt5 branch? > For Qt5, I do have -no-pch otherwise it doesn't build > (http://git.free-electrons.com/users/thomas-petazzoni/buildroot/commit/?h=qt5&id=7c1f7b9967d814e3f5c838e7090ac04b9d06af51). > Apparently, there are bugs in the precompiled headers implementation. Given that precompiled headers are disabled when using ccache, and they are broken otherwise (at least with some toolchains), I think it's best to just add -no-pch inconditionally. I'll send a patch that does so, and copy your comment at the commit you pointedme to: > -no-pch to workaround the issue described at > http://comments.gmane.org/gmane.comp.lib.qt.devel/5933 You might want to sign off the patch then. Luca