From mboxrd@z Thu Jan  1 00:00:00 1970
From: Paul Cercueil <paul@crapouillou.net>
Date: Fri, 20 Dec 2013 12:54:34 +0100
Subject: [Buildroot] [PATCH] cairo: Build with GLESv2 and VG backends if
	possible
In-Reply-To: <878uvgklnw.fsf@dell.be.48ers.dk>
References: <1387243053-5281-1-git-send-email-paul@crapouillou.net>
 <878uvgklnw.fsf@dell.be.48ers.dk>
Message-ID: <52B42FFA.2050202@crapouillou.net>
List-Id: <buildroot.busybox.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: buildroot@busybox.net

On 20/12/2013 08:48, Peter Korsgaard wrote:
>>>>>> "Paul" == Paul Cercueil <paul@crapouillou.net> writes:
>   > Signed-Off-By: Paul Cercueil <paul@crapouillou.net>
>   > ---
>   >  package/cairo/cairo.mk | 14 ++++++++++++++
>   >  1 file changed, 14 insertions(+)
>
>   > diff --git a/package/cairo/cairo.mk b/package/cairo/cairo.mk
>   > index 347bb75..aa03a27 100644
>   > --- a/package/cairo/cairo.mk
>   > +++ b/package/cairo/cairo.mk
>   > @@ -54,6 +54,20 @@ else
>   >  	CAIRO_CONF_OPT += --disable-directfb
>   >  endif
>   
>   > +ifeq ($(BR2_PACKAGE_HAS_OPENGL_ES),y)
>   > +	CAIRO_CONF_OPT += --enable-glesv2
>   > +	CAIRO_DEPENDENCIES += libgles
>   > +else
>   > +	CAIRO_CONF_OPT += --disable-glesv2
>   > +endif
>   > +
>   > +ifeq ($(BR2_PACKAGE_HAS_OPENVG),y)
>   > +	CAIRO_CONF_OPT += --enable-vg
>   > +	CAIRO_DEPENDENCIES += libopenvg
>   > +else
>   > +	CAIRO_CONF_OPT += --disable-vg
>   > +endif
>
> What openvg / gles providers did you test against? We seem to have a
> number of autobuild failures with the rpi provider:
>
> http://autobuild.buildroot.net/results/b7f/b7f71cd412e5a7f801e81140a02c7da04694151f/build-end.log
>
> http://autobuild.buildroot.net/results/f70/f7012c1e09d75dd103b61da364272f6385d2dc7b/build-end.log
>
> http://autobuild.buildroot.net/results/ec5/ec550547f6465b52ba552ca757c9bbeac17cd1c6/build-end.log
>
> http://autobuild.buildroot.net/results/81a/81ae93018ca6d7a61460e0c78d5a60e6a5531dbc/build-end.log
>
I am testing against Mesa.
I believe your target does not have Xorg, that would explain why 'None' 
is reported as undefined: it's defined in <X11/X.h>.
I can perfectly reproduce the issue if I disable Xorg. The source code 
uses 'None' even if targeting EGL and not GLX,
I can address that with another patch.

The second issue with 'GLchar' is surprising, since it's defined in 
<GLES2/gl2.h> here. I cannot even reproduce the issue.