[Buildroot] [autobuild.buildroot.net] Build results for 2013-05-29

[Thomas Petazzoni]

Dear Lionel Orry,

On Thu, 30 May 2013 09:27:33 +0200, Lionel Orry wrote:

> I noticed the error
> (http://autobuild.buildroot.net/results/2379e4c18d6b7ecc48741c53b3f2c2445cf5e75c/build-end.log)
> comes from the bundled 'src/cups/configure.in' which was generated by

I guess you meant src/cups/Makefile.in, right?

> a version of autotools which seems too old. The test of the existing
> directory in the resulting Makefile was missing $(DESTDIR) prefix
> while the $(MKDIR_P) afterwards was including it.
> 
> I executed ./autogen.sh and the result install code changed radically,
> $(MKDIR_P) not needing any directory existence test first.
> 
> I don't know what the procedure is in those case. I suppose we should
> suggest the gimp-print project devs to use a more recent autotools
> version for next release?

Yes, indeed.

> Are we allowed to call autogen.sh ourselves?

Preferably not autogen.sh. But you can add HOST_GUTENPRINT_AUTORECONF =
YES in gutenprint.mk to achieve the same goal. We're already doing an
autoreconf for the target build, so doing it on the host build
shouldn't be a problem.

Can you test this and see if it fixes the problem? If so submitting a
patch would be nice.

> Or provide a patch replacing the {configure,Makefile}.in files with
> more recent ones?

No, try with HOST_GUTENPRINT_AUTORECONF = YES. We don't want to carry
patches against configure/Makefile.in, as those patches are
generally really huge.

Also, if you're working on gutenprint, it would be nice if the
gutenprint-000-use-pregen-xmli18n-header.patch patch was rewritten to
use a new option in configure.ac, so that this patch could potentially
be upstreamed.

Best regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com