package/libva-intel-driver: Fix build

PATCH New Bernd Kuhls 3 participants 4 messages Mar 31, 2015 – Apr 4, 2015

lore

Share & tools

lore

Bernd Kuhls

author

2015-03-31 16:10 UTC #0 lore raw

PATCH 1/1 package/libva-intel-driver: Fix build

Package needs autoreconf to produce a working configure script, also fix a typo in .mk file

Signed-off-by Bernd Kuhls <[email protected]>

---
 package/libva-intel-driver/libva-intel-driver.mk |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

package/libva-intel-driver/libva-intel-driver.mk +2 -1

index 5755207..560ca28 100644
@@ -9,6 +9,7 @@ LIBVA_INTEL_DRIVER_SOURCE = libva-intel-driver-$(LIBVA_INTEL_DRIVER_VERSION).tar
 LIBVA_INTEL_DRIVER_SITE = http://www.freedesktop.org/software/vaapi/releases/libva-intel-driver
 LIBVA_INTEL_DRIVER_LICENSE = MIT
 LIBVA_INTEL_DRIVER_LICENSE_FILES = COPYING
+LIBVA_INTEL_DRIVER_AUTORECONF = YES
 LIBVA_INTEL_DRIVER_DEPENDENCIES = host-pkgconf libdrm libva
 
 ifeq ($(BR2_PACKAGE_XORG7),y)

@@ -19,7 +20,7 @@ LIBVA_INTEL_DRIVER_CONF_OPTS += --disable-x11
 endif
 
 ifeq ($(BR2_PACKAGE_WAYLAND),y)
-LIBVA_DEPENDENCIES += wayland
+LIBVA_INTEL_DRIVER_DEPENDENCIES += wayland
 LIBVA_INTEL_DRIVER_CONF_OPTS += --enable-wayland
 else
 LIBVA_INTEL_DRIVER_CONF_OPTS += --disable-wayland

Signature · 1 line

1.7.10.4

Thomas Petazzoni

2015-03-31 21:40 UTC #1 lore raw

PATCH 1/1 package/libva-intel-driver: Fix build

Dear Bernd Kuhls,

Quoted 23 lines · Bernd Kuhls

Package needs autoreconf to produce a working configure script, also fix a typo in .mk file
Fixes
http://autobuild.buildroot.net/results/606/6061216fa6e2f4323117bb6cfb2b77753a89a679/
http://autobuild.buildroot.net/results/76d/76dc4b670e2131b94bd0c5cff45ccafd5e45033c/
http://autobuild.buildroot.net/results/d6f/d6f624baa3dde4fc4051d361ddaf98ecc5c84134/
http://autobuild.buildroot.net/results/91f/91f36f32ef43be897fef9352f97cc07781c4035a/
http://autobuild.buildroot.net/results/951/9517bb6d2c05c27cad0b24a4443021a34babedbd/
http://autobuild.buildroot.net/results/405/405dfffbcfdaa485d0000dad655b0c112688a58f/
http://autobuild.buildroot.net/results/925/92538fb07d5226bb8c3d3cc5e1429560b2e621f3/
http://autobuild.buildroot.net/results/d15/d154ca018da0c21762ed70acda4f224c50a1d808/
Signed-off-by Bernd Kuhls <[email protected]>
---
 package/libva-intel-driver/libva-intel-driver.mk |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
package/libva-intel-driver/libva-intel-driver.mk +0 -0
index 5755207..560ca28 100644

package/libva-intel-driver/libva-intel-driver.mk LIBVA_INTEL_DRIVER_SOURCE = libva-intel-driver-$(LIBVA_INTEL_DRIVER_VERSION).tar · line 12

Quoted 4 lines

package/libva-intel-driver/libva-intel-driver.mk +1 -0

@@ -9,6 +9,7 @@ LIBVA_INTEL_DRIVER_SOURCE = libva-intel-driver-$(LIBVA_INTEL_DRIVER_VERSION).tar
 LIBVA_INTEL_DRIVER_SITE = http://www.freedesktop.org/software/vaapi/releases/libva-intel-driver
 LIBVA_INTEL_DRIVER_LICENSE = MIT
 LIBVA_INTEL_DRIVER_LICENSE_FILES = COPYING
+LIBVA_INTEL_DRIVER_AUTORECONF = YES

Can you comment a bit more on why this is needed? Your commit lo says "needed to produce a working configure script", but a quick look at the build failures only points to the wayland dependency problem.

Thanks,

Thomas

Signature · 3 lines

Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

Bernd Kuhls

author

2015-04-02 20:12 UTC #2 lore raw

PATCH 1/1 package/libva-intel-driver: Fix build

Thomas Petazzoni <thomas.petazzoni-wi1+55ScJUtKEb57/3fJTNBPR1lH4CV8 @public.gmane.org> wrote in news:20150331234038.5fe3b78b at free-electrons.com:

Quoted 3 lines

Can you comment a bit more on why this is needed? Your commit lo says "needed to produce a working configure script", but a quick look at the build failures only points to the wayland dependency problem.

Hi,

when you compare the configure script delivered by libva-intel-driver and the one generated by our autoreconf you will see some huge differences.

The upstream configure script does not honour --disable-wayland, instead it always assumes wayland support for a reason I do not know. Maybe the line

m4_ifdef([WAYLAND_SCANNER_RULES],

in configure.ac, line 170, is not interpreted correctly by the autoconf tools used upstream. Running autoreconf in buildroot fixes the build error.

Regards, Bernd

Peter Korsgaard

2015-04-04 22:03 UTC #3 lore raw

PATCH 1/1 package/libva-intel-driver: Fix build

Quoted 14 lines

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Bernd" == Bernd Kuhls <[email protected]> writes:
Thomas Petazzoni <thomas.petazzoni-wi1+55ScJUtKEb57/3fJTNBPR1lH4CV8 @public.gmane.org> wrote in news:20150331234038.5fe3b78b at free-electrons.com:
Quoted 3 lines
Can you comment a bit more on why this is needed? Your commit lo says "needed to produce a working configure script", but a quick look at the build failures only points to the wayland dependency problem.
Hi,
when you compare the configure script delivered by libva-intel-driver and the one generated by our autoreconf you will see some huge differences.

Quoted 7 lines

The upstream configure script does not honour --disable-wayland, instead it always assumes wayland support for a reason I do not know. Maybe the line
m4_ifdef([WAYLAND_SCANNER_RULES],
in configure.ac, line 170, is not interpreted correctly by the autoconf tools used upstream. Running autoreconf in buildroot fixes the build error.

Ahh, I get it now.

This is not really about a wrongly generated configure script, but simply because the WAYLAND_SCANNER handling is outside the wayland conditional.

There is an open bugzilla entry about it upstream:

https://bugs.freedesktop.org/show_bug.cgi?id=79478

WAYLAND_SCANNER_RULES is provided by wayland, so the reason why running autoreconf fixes the issue is that it will only get expanded if built after wayland.

Committed with a more detailed commit message, thanks.

Signature · 1 line

Bye, Peter Korsgaard