[1/2] package/webkit: Fix detection of egl headers provided by rpi-userland

quote from webkit/config.log:
/home/br2/output/host/usr/arm-buildroot-linux-uclibcgnueabi/sysroot/usr/include/interface/vcos/vcos_types.h:38:33:
fatal error: vcos_platform_types.h: No such file or directory

Package rpi-userland provides this header file, but in a non-default directory,
therefore use pkg-config to add the appropriate flags.

Fixes
http://autobuild.buildroot.net/results/bc0/bc0f9114f9dacf68a510f1e01f2dc3661c34f536/

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
---
 package/webkit/webkit.mk |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

Dear Bernd Kuhls,

On Thu, 29 May 2014 14:26:32 +0200, Bernd Kuhls wrote:
> quote from webkit/config.log:
> /home/br2/output/host/usr/arm-buildroot-linux-uclibcgnueabi/sysroot/usr/include/interface/vcos/vcos_types.h:38:33:
> fatal error: vcos_platform_types.h: No such file or directory
> 
> Package rpi-userland provides this header file, but in a non-default directory,
> therefore use pkg-config to add the appropriate flags.
> 
> Fixes
> http://autobuild.buildroot.net/results/bc0/bc0f9114f9dacf68a510f1e01f2dc3661c34f536/
> 
> Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
> ---
>  package/webkit/webkit.mk |    6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)

Could you update on whether this patch and patch 2/2 in this series are
still needed? There were also unanswered questions in your patch 2/2.

Thanks,

Thomas

Thomas Petazzoni <thomas.petazzoni-wi1+55ScJUtKEb57/3fJTNBPR1lH4CV8
@public.gmane.org> wrote in news:20140723235453.5ebea556@free-electrons.com:

> Could you update on whether this patch and patch 2/2 in this series are
> still needed? There were also unanswered questions in your patch 2/2.

Hi Thomas,

the autobuild error I was trying to fix occured on May 29th:
http://autobuild.buildroot.net/results/bc0/bc0f9114f9dacf68a510f1e01f2dc3661c
34f536/

The same error was still present on July, 21th:
http://autobuild.buildroot.net/results/1e8/1e8cab75025316b22205be7fb08a93148a
256b9c/

As I wrote in my cover letter[1] for this patch series

> the rpi experts should have a closer look whether webkit could work on rpi
> despite the current problem

I asked for this help because I have basically no knowledge about the rpi-
specific binary blobs provided by rpi-userland.

Maybe Hadrien´s patch series[2] to bump webkit to 2.4.3 will solve the 
problem in a better way than with my patches.

@Hadrien: Could you please test your patch using the defconfig used by the 
autobuilder where webkit-1 failed to compile for rpi?

Regards, Bernd

[1] http://article.gmane.org/gmane.comp.lib.uclibc.buildroot/85281
[2] http://thread.gmane.org/gmane.comp.lib.uclibc.buildroot/89775

Hi Bernd,

On Fri, 25 Jul 2014 21:34:07 +0200, Bernd Kuhls wrote:
> Thomas Petazzoni <thomas.petazzoni-wi1+55ScJUtKEb57/3fJTNBPR1lH4CV8
> @public.gmane.org> wrote in news:20140723235453.5ebea556@free-electrons.com:
> 
>> Could you update on whether this patch and patch 2/2 in this series are
>> still needed? There were also unanswered questions in your patch 2/2.
> 
> Hi Thomas,
> 
> the autobuild error I was trying to fix occured on May 29th:
> http://autobuild.buildroot.net/results/bc0/bc0f9114f9dacf68a510f1e01f2dc3661c
> 34f536/
> 
> The same error was still present on July, 21th:
> http://autobuild.buildroot.net/results/1e8/1e8cab75025316b22205be7fb08a93148a
> 256b9c/
> 
> As I wrote in my cover letter[1] for this patch series
> 
>> the rpi experts should have a closer look whether webkit could work on rpi
>> despite the current problem
> 
> I asked for this help because I have basically no knowledge about the rpi-
> specific binary blobs provided by rpi-userland.
> 
> Maybe Hadrien´s patch series[2] to bump webkit to 2.4.3 will solve the 
> problem in a better way than with my patches.
> 
> @Hadrien: Could you please test your patch using the defconfig used by the 
> autobuilder where webkit-1 failed to compile for rpi?

Actually, I had forgotten this problem but yes, when I initially tested
the bump of WebKit with rpi-userland the EGL check failed too. So it's
not solved in my series or with version 2.4.3 and this seems to be
related to rpi-userland (no problem with ti-gfx and imx-lib, with which
I test now).

I will test my patch with the defconfig used by the autobuilder and test
your patch if the problem occurs again.

Regards,

Hadrien