From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Thu, 23 Oct 2014 22:08:56 +0200 Subject: [Buildroot] [PATCH] intltool: fix build In-Reply-To: References: <1413571681-27963-1-git-send-email-francois.perrad@gadz.org> <20141022180339.47298e90@free-electrons.com> Message-ID: <20141023220856.2bd3203d@free-electrons.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Dear Fran?ois Perrad, On Wed, 22 Oct 2014 19:15:36 +0200, Fran?ois Perrad wrote: > > Can you cook an updated version of this patch that takes into account > > the comment you have received? It would help us fix a number of Perl > > related autobuilder issues. > > When I try to do that you suggest, I only obtain no working result. > > I think that comments from Arnout, come from a pure review but no from > testing or experiment. > > This patch doesn't try to fix the package host-intltool, it fixes the > packages which depend on host-intltool. > Currently, this kind of package is broken in the configure step > because they can not find the correct perl. > So, the variable INTLTOOL_PERL=$(PERL) in *_CONFIGURE_OPTS is mandatory. > > When PATH=$(BR_PATH), `which perl` find the system perl when host-perl > is not yet installed, after it find the host-perl. > So, it is important to save the result of `which perl` with the > original PATH (ie. without $(HOST_DIR)/usr/bin). > > The current patch stay the best that I can do (the commit message can > always be improved). Hum, right, ok. I'm not a big fan of how we handle intltool, but I admit that the main reason for this mess is me asking you to not make intltool and libxml-parser-perl depend on host-perl. I've applied your patch (after slightly rewording the commit log), it indeed seems to fix the issue. We'll see how things go, but I'm not very happy with our intltool handling :/ Best regards, Thomas -- Thomas Petazzoni, CTO, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com