]> code.ossystems Code Review - openembedded-core.git/log
openembedded-core.git
13 years agoxf86-video-intel: Update to 2.17.0
Saul Wold [Tue, 29 Nov 2011 19:41:21 +0000 (11:41 -0800)]
xf86-video-intel: Update to 2.17.0

* Remove patch for issue fixed upstream

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoxf86-input-synaptics: Update to 1.5.0
Saul Wold [Tue, 29 Nov 2011 19:40:41 +0000 (11:40 -0800)]
xf86-input-synaptics: Update to 1.5.0

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoxf86-input-mouse: Update to 1.7.1
Saul Wold [Tue, 29 Nov 2011 19:40:12 +0000 (11:40 -0800)]
xf86-input-mouse: Update to 1.7.1

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoxwininfo: Update to 1.1.2
Saul Wold [Tue, 29 Nov 2011 19:39:43 +0000 (11:39 -0800)]
xwininfo: Update to 1.1.2

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoxset: Update to 1.2.2
Saul Wold [Tue, 29 Nov 2011 19:39:21 +0000 (11:39 -0800)]
xset: Update to 1.2.2

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoxrandr: Update to 1.3.5
Saul Wold [Tue, 29 Nov 2011 19:39:02 +0000 (11:39 -0800)]
xrandr: Update to 1.3.5

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoxprop: Update to 1.2.1
Saul Wold [Tue, 29 Nov 2011 19:38:26 +0000 (11:38 -0800)]
xprop: Update to 1.2.1

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoxkbcomp: Update to 1.2.3
Saul Wold [Tue, 29 Nov 2011 19:38:06 +0000 (11:38 -0800)]
xkbcomp: Update to 1.2.3

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoxinit: Update to 1.3.1
Saul Wold [Tue, 29 Nov 2011 19:37:30 +0000 (11:37 -0800)]
xinit: Update to 1.3.1

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoxdpyinfo: Update to 1.3.0
Saul Wold [Tue, 29 Nov 2011 19:14:11 +0000 (11:14 -0800)]
xdpyinfo: Update to 1.3.0

* Rebase patch to update code

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agomkfontscale: Update to 1.0.9
Saul Wold [Tue, 29 Nov 2011 19:13:05 +0000 (11:13 -0800)]
mkfontscale: Update to 1.0.9

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agolibdrm: Update to 2.4.27
Saul Wold [Tue, 29 Nov 2011 19:11:59 +0000 (11:11 -0800)]
libdrm: Update to 2.4.27

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agojson-glib: Update to 0.14.2
Saul Wold [Tue, 29 Nov 2011 19:10:52 +0000 (11:10 -0800)]
json-glib: Update to 0.14.2

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agowget: Update to 1.13.4
Saul Wold [Tue, 29 Nov 2011 19:02:41 +0000 (11:02 -0800)]
wget: Update to 1.13.4

* remove gnutls.bzr patch as it was in upstream

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoqt-mobility: Add missing patch
Richard Purdie [Tue, 29 Nov 2011 14:36:09 +0000 (14:36 +0000)]
qt-mobility: Add missing patch

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agolibsdl: enable alsa based on PACKAGECONFIG and respect DISTRO_FEATURES
Martin Jansa [Wed, 9 Nov 2011 15:28:19 +0000 (16:28 +0100)]
libsdl: enable alsa based on PACKAGECONFIG and respect DISTRO_FEATURES

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
13 years agoqt4: qt-mobility: fix QML video player crash
Anatolij Gustschin [Tue, 29 Nov 2011 13:12:47 +0000 (14:12 +0100)]
qt4: qt-mobility: fix QML video player crash

Fix QML video player crash which happens when attempting
to play OGG video files.

Signed-off-by: Anatolij Gustschin <agust@denx.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agoqt4: fix sed to have correct *.pc files
Eric Bénard [Tue, 29 Nov 2011 10:47:05 +0000 (11:47 +0100)]
qt4: fix sed to have correct *.pc files

it seems sed doesn't handle \? properly if the string in not between
quotes.

without this patch, we get something like (for example for QtDBusE.pc :
Libs: -L${libdir} -lQtDBusE
Libs.private: -L/home/ebenard/WORK/setup-scripts/build/tmp-angstrom_2010_x-eglibc/sysroots/board/
usr/lib -L/usr/lib  -L/home/ebenard/WORK/setup-scripts/build/tmp-angstrom_2010_x-eglibc/sysroots/board
/usr/lib -lQtXmlE -L/usr/lib -lQtCoreE -lpthread
Cflags: -DQT_SHARED -I/usr/include/qtopia -I${includedir}

with the patch we get what is expected :
Libs: -L${libdir} -lQtDBusE
Libs.private:     -lQtXmlE  -lQtCoreE -lpthread
Cflags: -DQT_SHARED -I/usr/include/qtopia -I${includedir}

Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agolibsdl: replace tabs with spaces
Martin Jansa [Tue, 15 Nov 2011 23:14:15 +0000 (00:14 +0100)]
libsdl: replace tabs with spaces

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
13 years agolibsdl: enable cdrom, alsa and tslib, disable rpath and add few fixes from meta-oe
Martin Jansa [Fri, 4 Nov 2011 16:19:35 +0000 (17:19 +0100)]
libsdl: enable cdrom, alsa and tslib, disable rpath and add few fixes from meta-oe

* virtual/libsdl is used in sdl.bbclass but nothing was providing it
* rpath disabled because otherwise libtool and ld try to link against
  ${libdir}/libdl.so.2 on the host
  (see OE-classic commit af1a2e0e7626e372f22afbcabf08d9ae6b0d7b01)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
13 years agolibsdl: rename files dir to libsdl-1.2.14 for faster lookup
Martin Jansa [Wed, 9 Nov 2011 15:23:55 +0000 (16:23 +0100)]
libsdl: rename files dir to libsdl-1.2.14 for faster lookup

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
13 years agolibsdl: drop unused files
Martin Jansa [Wed, 9 Nov 2011 15:23:20 +0000 (16:23 +0100)]
libsdl: drop unused files

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
13 years agoeglibc-initial: Ensure symlinks point to the correct location when built from sstate...
Richard Purdie [Tue, 29 Nov 2011 12:21:07 +0000 (12:21 +0000)]
eglibc-initial: Ensure symlinks point to the correct location when built from sstate cache

If the sstate files are installed into a sysroot from the sstate cache,
the directory to the main sysroot can change and the symlinks aren't
adjusted to account for this. This is a problem specific to the toolchain
bootstrap process. This patch adds up a function to recreate the
symlinks, hence ensuring they always point at the correct location.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agomesa-common: allow empty PN
Martin Jansa [Fri, 25 Nov 2011 08:18:56 +0000 (09:18 +0100)]
mesa-common: allow empty PN

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
13 years agobase-passwd: Fix race
Richard Purdie [Tue, 29 Nov 2011 10:40:23 +0000 (10:40 +0000)]
base-passwd: Fix race

The do_package task this recipe injects depends on the output of
do_populate_sysroot to be present. This introduces the correct
dependency so that the package task works correctly.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agodistro-tracking: update information for binutils & libtool
Nitin A Kamble [Mon, 28 Nov 2011 18:22:05 +0000 (10:22 -0800)]
distro-tracking: update information for binutils & libtool

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
13 years agobinutils: upgrade from 2.21.1a to 2.22
Nitin A Kamble [Mon, 28 Nov 2011 18:20:10 +0000 (10:20 -0800)]
binutils: upgrade from 2.21.1a to 2.22

And rebase the patches to the newer source code

This patch is upstream hence deleting it from the recipe.
binutils/110-arm-eabi-conf.patch

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
13 years agodistro_tracking: update tracking fields for gettext and eds-tools
Wenzong Fan [Fri, 25 Nov 2011 06:59:11 +0000 (14:59 +0800)]
distro_tracking: update tracking fields for gettext and eds-tools

Add maintainer and update manual_check_date, status fields.

Signed-off-by: Wenzong Fan <wenzong.fan@windriver.com>
13 years agoSelf Hosted Image: Added Task and Images
Saul Wold [Thu, 3 Nov 2011 20:11:41 +0000 (13:11 -0700)]
Self Hosted Image: Added Task and Images

This is the first pass of creating a self hosted image, this task
and image can pass bitbake's sanity check (if modified to remove the
cvs check) and can build psuedo, albit very slowly on qemu.

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
13 years agointltool: remove XML::Parser check
Saul Wold [Wed, 23 Nov 2011 23:47:27 +0000 (15:47 -0800)]
intltool: remove XML::Parser check

Add Patch to disable the XML::Parser check in the target
intltool.m4, this check will find the host (not native)
XML::Parser if it's installed possibly causing Host
contamination, but will also fail configuration if XML::Parser
is not installed on the host.

Since we know that XML::Parser is installed on the image, we don't
really need this check, so comment it out.

From RP in mail thread:
> If the recipe needs perl for
> some other reason than intltool, it needs perlnative but it if only
> needs perl for intltool, we shouldn't need the dependency. The .m4 macro
> checks are well intended but don't fit the way we use perl. I really
> don't want to end up in a position where intltool automatically means we
> have to add perlnative as a dependency and we've previously seen many
> problems related to that.

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agotzdata/tzcode: Update tzdata to 2011n
Saul Wold [Mon, 28 Nov 2011 18:37:56 +0000 (10:37 -0800)]
tzdata/tzcode: Update tzdata to 2011n

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agomesa: package gl/egl/osmesa to separate packages
Martin Jansa [Thu, 13 Oct 2011 08:07:57 +0000 (10:07 +0200)]
mesa: package gl/egl/osmesa to separate packages

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
13 years agodistro_tracking_fields: update the status
Xiaofeng Yan [Thu, 24 Nov 2011 07:47:13 +0000 (15:47 +0800)]
distro_tracking_fields: update the status

directfb to 1.4.15
xorg-xserver-lite to 1.11.1
libx11-diet to 1.4.4

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
13 years agodistro_tracking: Update manual check dates (readline, gcrypt)
Saul Wold [Thu, 24 Nov 2011 00:41:19 +0000 (16:41 -0800)]
distro_tracking: Update manual check dates (readline, gcrypt)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agorunqemu: look for *-image* to be more flexible for image names
Saul Wold [Fri, 4 Nov 2011 00:37:30 +0000 (17:37 -0700)]
runqemu: look for *-image* to be more flexible for image names

We can't just look for *image* since it will catch up bzimage

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agoxserver-xorg: only use --disable-dga in xserver-xorg-lite
Otavio Salvador [Wed, 23 Nov 2011 18:50:20 +0000 (18:50 +0000)]
xserver-xorg: only use --disable-dga in xserver-xorg-lite

Xorg VESA driver calls DGAInit enforcing this symbol to be available
to the driver to work so we revert back to only disable DGA for lite
flavor of Xorg.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
13 years agolibx11-trim: update to 1.4.4
Xiaofeng Yan [Mon, 21 Nov 2011 07:42:24 +0000 (15:42 +0800)]
libx11-trim: update to 1.4.4

Some description :
1 Remove --with-xcb because xcb is necessary item in the new version.
2 keysymdef_include.patch uses in keeping native directory of X11 as the default configuration. \
  or else host directory "/usr/include/X11" is selected when running  ./configure.
3 makekeys_crosscompile.patch avoid host contamination.
4 x11_disable_makekeys.patch uses in compiling makekey.c needed by makekeys-makekeys.o

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
13 years agorpm_5.4.0.bb: Build rpm without xz
Matthew McClintock [Mon, 28 Nov 2011 21:22:47 +0000 (15:22 -0600)]
rpm_5.4.0.bb: Build rpm without xz

This fixes the following issue:

Log data follows:
| NOTE: Creating RPM package for perf-dbg
| NOTE: Creating RPM package for perf
| NOTE: Creating EMPTY RPM Package for kernel
| NOTE: Creating EMPTY RPM Package for kernel-3.0.9-00348-gec4b357
| NOTE: Creating RPM package for kernel-image-3.0.9-00348-gec4b357
| NOTE: Creating RPM package for kernel-dev
| NOTE: Creating RPM package for kernel-vmlinux
| NOTE: Not creating empty RPM package for kernel-misc
| NOTE: Creating RPM package for kernel-devicetree
| NOTE: Creating RPM package for kernel-module-libcrc32c
| NOTE: Creating RPM package for kernel-module-crc-itu-t
| NOTE: Creating RPM package for kernel-module-sctp
| NOTE: Creating RPM package for kernel-module-pcbc
| NOTE: Creating RPM package for kernel-module-crc32c
| NOTE: Creating RPM package for kernel-module-binfmt-misc
| NOTE: Creating RPM package for kernel-module-nfsd
| NOTE: Creating RPM package for kernel-module-exportfs
| NOTE: Creating RPM package for kernel-module-msdos
| NOTE: Creating RPM package for kernel-module-nls-utf8
| NOTE: Creating RPM package for kernel-module-udf
| NOTE: Creating RPM package for kernel-module-isofs
| NOTE: Creating RPM package for kernel-module-usbhid
| NOTE: Creating RPM package for kernel-module-scsi-wait-scan
| NOTE: Creating EMPTY RPM Package for kernel-modules
| /local/home/mattsm/git/fsl-local-sdk/build_p4080ds_release/tmp/sysroots/x86_64-linux/usr/bin/rpmbuild.real: error while loading shared libraries: liblzma.so.5: cannot open shared object file: No such file or directory
| ERROR: Function 'BUILDSPEC' failed (see /local/home/mattsm/git/fsl-local-sdk/build_p4080ds_release/tmp/work/p4080ds-fsl-linux/linux-qoriq-sdk-3.0.6-r2/temp/log.do_package_write_rpm.18943 for further information)

Signed-off-by: Matthew McClintock <msm@freescale.com>
13 years agoIntroduce multiarch DISTRO_FEATURE
Julian Pidancet [Wed, 16 Nov 2011 01:01:59 +0000 (01:01 +0000)]
Introduce multiarch DISTRO_FEATURE

This patch introduces a distro feature which enables gcc to produce
both 32bit and 64bit code, and enables binutils to operate on both
32bit and 64bit binaries. It differs from multilib toolchains in
that it does not require to compile a version of the libc for each
architecture variant. However, the code produced for the secondary
architecture will not be linkable against the libc.

v2: - Renamed the feature name from "biarch" to "multiarch". The GCC
installation manual claims that the mips-linux can be made a tri-arch
compiler (http://gcc.gnu.org/install/configure.html)
    - For x86_64, the compiler is made bi-arch by default, so nothing
has to be done in particular.
    - I analyzed the gcc/config.gcc from GCC sources and added in this
patch all the architectures that could be made biarch with the version
of gcc currently used in OE, which are powerpc, and sparc, in addition
to x86. mips and s390 will probably be supported in future versions of
gcc. For x86 and sparc, only the --enable-targets=all option is valid
to make this work (this option doesn't have any other side effects than
making the compiler bi-arch). For powerpc, I used the
--enable-targets=powerpc64 option (although 'all' also works).

Note: - Untested on powerpc and sparc. But I believe it works the same
as with x86.
      - gcc in meta-toolchain is also made multiarch.

Signed-off-by: Julian Pidancet <julian.pidancet@gmail.com>
13 years agoopkg-utils: handle 6 digit UID fields properly
Koen Kooi [Sat, 26 Nov 2011 09:47:29 +0000 (10:47 +0100)]
opkg-utils: handle 6 digit UID fields properly

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
13 years agopython-native: Fix gcc compiler detecting logic
Khem Raj [Mon, 28 Nov 2011 01:22:37 +0000 (17:22 -0800)]
python-native: Fix gcc compiler detecting logic

We have a patch unixccompiler.patch where we try to throw away
everything except first element of CC string but this does not
work if gcc is prepended with something e.g. CC="ccache gcc"
then the logic fails and it ends up in some modules failing on
you silently (_sqlite3) in my case.
The fix here is to drop basename so we keep the whole
string as it is and then the detection function searches
for gcc string in the whole CC. This works in both cases
one the original intent of the patch and the second described
above. One place where it will fail is if someone has non-gcc
compiler installed in some subdir which has gcc in it e.g.
/usr/gcc/fakecc but for OE this should never happen. Ideally
the the detection logic should have tried to execute gcc
and then parsed --version output or something.

Signed-off-by: Khem Raj <raj.khem@gmail.com>
13 years agotoolchain-scripts.bbclass: Make it work when TCLIBC=uclibc
Khem Raj [Mon, 28 Nov 2011 01:29:24 +0000 (17:29 -0800)]
toolchain-scripts.bbclass: Make it work when TCLIBC=uclibc

This class currently only works with eglibc. Since
it adds dependencies explicitly on eglibc when using
uclibc this creates problems. So we make sure that
it checks for TCLIBC to determine system C library
in use

Signed-off-by: Khem Raj <raj.khem@gmail.com>
13 years agodefault-distrovars: Define SDK_VERSION and DISTRO_VERSION
Khem Raj [Mon, 28 Nov 2011 01:32:00 +0000 (17:32 -0800)]
default-distrovars: Define SDK_VERSION and DISTRO_VERSION

runqemu scripts check for them and when using just oe-core
these are not defined anywhere

Signed-off-by: Khem Raj <raj.khem@gmail.com>
13 years agopulseaudio: update to 1.1, delete 0.9.x
Koen Kooi [Thu, 24 Nov 2011 09:18:12 +0000 (10:18 +0100)]
pulseaudio: update to 1.1, delete 0.9.x

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
13 years agolibjson: add 0.9
Koen Kooi [Thu, 24 Nov 2011 09:18:11 +0000 (10:18 +0100)]
libjson: add 0.9

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
13 years agogetVar/setVar cleanups
Richard Purdie [Fri, 25 Nov 2011 14:25:16 +0000 (14:25 +0000)]
getVar/setVar cleanups

Complete the bb.data.getVar/setVar replacements with accesses
directly to the data store object.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agoghostscript: Ensure we run reautoconf
Richard Purdie [Fri, 25 Nov 2011 13:15:04 +0000 (13:15 +0000)]
ghostscript: Ensure we run reautoconf

Avoding the autoreconf with a hardcoded do_configure is bad practise
since it can hide various errors. This patch ensures we do use the
standard do_configure.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agoghostscript-native: Ensure the sys/time/h fix is applied for native builds
Richard Purdie [Fri, 25 Nov 2011 13:13:30 +0000 (13:13 +0000)]
ghostscript-native: Ensure the sys/time/h fix is applied for native builds

On my system, the sys/time.h header is in a subdir off /usr/include
which causes a build failure. Apply the target CFLAGS fix to native
builds as well to address this.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agoghostscript: Really fix parallel make race
Richard Purdie [Fri, 25 Nov 2011 13:10:41 +0000 (13:10 +0000)]
ghostscript: Really fix parallel make race

There continue to be parallel make race issues showing up on the autobuilder.
This patch removes some potential sources of these. The rm is unrequired
since we're using cp -f. The || true ensures that if we did race against
someone it becomes harmless.

[YOCTO #1202]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agoofono: Fix missing python dependency and package tests separately
Richard Purdie [Thu, 24 Nov 2011 23:04:13 +0000 (23:04 +0000)]
ofono: Fix missing python dependency and package tests separately

We need to ensure there is a python dependency for the ofono python
test scripts. To do this without introducing the dependency
unnecessarily for the core ofono functionality, this patch splits
the tests into a separate package too.

[YOCTO #1705]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agoNothing uses USERNAME, remove it - can cause sstate-cache conflicts
Matthew McClintock [Thu, 24 Nov 2011 01:12:31 +0000 (19:12 -0600)]
Nothing uses USERNAME, remove it - can cause sstate-cache conflicts

USER is the correct variable to use, also this can affect sstate
cache as well.

Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agogcc-package-target.inc: add the symbol link /lib/cpp
Dexuan Cui [Wed, 16 Nov 2011 09:49:44 +0000 (17:49 +0800)]
gcc-package-target.inc: add the symbol link /lib/cpp

When I was trying self-hosted-image, eglibc's do_install failed in the target:
ERROR: cannot stat bootparam_prot.h:
the cause is: rpcgen doesn't work properly: rpcgen can't exec /lib/cpp since
it doesn't exist.

According to http://tldp.org/LDP/Linux-Filesystem-Hierarchy/html/lib.html:
"if a C preprocessor is installed, /lib/cpp must be a reference to it, for
historical reasons. The usual placement of this binary is /usr/bin/cpp".

Typical distros, like Ubuntu, openSuSE, Fedora and RHEL, all comply with
the rule.

Actually in meta/recipes-devtools/gcc/gcc-package-target.inc, we do try to
package ${base_libdir}/cpp:
 FILES_cpp = "\
  ${bindir}/${TARGET_PREFIX}cpp \
  ${base_libdir}/cpp \
  ${libexecdir}/gcc/${TARGET_SYS}/${BINV}/cc1"
But unluckily we didn't create a symbol link in do_install.
This patch adds the symbol link.

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
13 years agosite/x86_64-linux: add cvs config variables
Nitin A Kamble [Wed, 23 Nov 2011 05:55:51 +0000 (21:55 -0800)]
site/x86_64-linux: add cvs config variables

configure of cvs packages was failing on the meta-toolchain for a x86_64 target.

Configure error reported:
checking whether printf supports %p... configure: error: cannot run test program while cross compiling

This fixes [YOCTO #1781]

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agobitbake.conf: We only care about the absolute value of baselib
Richard Purdie [Thu, 24 Nov 2011 17:46:46 +0000 (17:46 +0000)]
bitbake.conf: We only care about the absolute value of baselib

The value of baselib can be constructed in several different ways
and from a sstate perspective we don't care how it was made up,
we only care what the final value is. This uses the new functionality
in bitbake to ensure we only include the value of baselib and not
any intermediate dependencies.

[YOCTO #1583]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agoavahi-ui: Fix packaging and conflict with the avahi recipe
Richard Purdie [Thu, 24 Nov 2011 15:17:58 +0000 (15:17 +0000)]
avahi-ui: Fix packaging and conflict with the avahi recipe

We need to remove the files from avahi-ui which conflict with the avahi recipe.
If we don't do this they trigger packaging warnings and can also overwrite
files in the sysroot unexpectedly causing build failures (if X depends on
avahi, it expects avahi's files to stay there, not disappear as avahi-ui
build/stages).

This patch cleans up the packaging although I wish there were an alterantive
to the do_install which makes my eyes bleed.

[YOCTO #1770]
[YOCTO #1722]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agoeglibc.inc: Remove EGLIBPARALLELISM from deps for EXTRA_OEMAKE
Matthew McClintock [Thu, 17 Nov 2011 22:28:20 +0000 (16:28 -0600)]
eglibc.inc: Remove EGLIBPARALLELISM from deps for EXTRA_OEMAKE

Without this simply changing the number of threads via
PARALLEL_MAKE can invalidate sstate-cache

Signed-off-by: Matthew McClintock <msm@freescale.com>
13 years agopatch.bbclass: Add PATCHRESOLVE to excluded vars for generating sstate-cache
Matthew McClintock [Thu, 17 Nov 2011 22:42:47 +0000 (16:42 -0600)]
patch.bbclass: Add PATCHRESOLVE to excluded vars for generating sstate-cache

The method of resolving the patch should not effect the sstate-cache
signature.

Signed-off-by: Matthew McClintock <msm@freescale.com>
13 years agogtk+: Have gtk+ pick up ${NM} from the environment
Xiaofeng Yan [Thu, 17 Nov 2011 10:36:50 +0000 (18:36 +0800)]
gtk+: Have gtk+ pick up ${NM} from the environment

when gtk+ configures, it should pick up ${NM} from the environment \
if it is defined, instead of just looking for "nm".

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
13 years agotslib: Set open mode for ts_calibate.c
Xiaofeng Yan [Thu, 17 Nov 2011 11:03:21 +0000 (19:03 +0800)]
tslib: Set open mode for ts_calibate.c

Get patch from: http://permalink.gmane.org/gmane.comp.lib.uclibc.buildroot/27614

There is a potentially "unsafe" use of open(). Ubuntu 8.10+, for instance,
by default enables the compiler flag -D_FORTIFY_SOURCE=2 \
which throws an error of ts_calibrate.c. To fix this, \
set a mode in the open() call, patch patched ts_calibrate.c \
to set 0644 (S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH) and it compiles fine

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
13 years agorpm: Flush old logs by change the DB_CONFIG
Mei Lei [Thu, 17 Nov 2011 06:57:56 +0000 (14:57 +0800)]
rpm: Flush old logs by change the DB_CONFIG

Fixes [YOCTO #1174]

Rpm logs will grow indefinitely, so change the config to flush those old logs.

Signed-off-by: Mei Lei <lei.mei@intel.com>
13 years agoeglibc_2.13: update SRCREV
Kang Kai [Wed, 16 Nov 2011 07:50:23 +0000 (15:50 +0800)]
eglibc_2.13: update SRCREV

Update eglibc 2.13 SRCREV to 15508, just to sync with upstream
Remove glibc_bug_fix_12454.patch, because it is already merged.

Signed-off-by: Kang Kai <kai.kang@windriver.com>
13 years agobase-passwd: add Upstream-Status to patches, remove unused patch
Scott Garman [Tue, 15 Nov 2011 01:11:49 +0000 (17:11 -0800)]
base-passwd: add Upstream-Status to patches, remove unused patch

mysql.patch is not used by the recipe anymore, so delete it.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
13 years agodosfstools: add Upstream-Status to patches
Scott Garman [Tue, 15 Nov 2011 01:03:41 +0000 (17:03 -0800)]
dosfstools: add Upstream-Status to patches

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
13 years agobitbake.conf: Set BB_CONSOLELOG so we log console messages by default
Richard Purdie [Thu, 24 Nov 2011 14:50:37 +0000 (14:50 +0000)]
bitbake.conf: Set BB_CONSOLELOG so we log console messages by default

The console log is often helpful for debugging issues. Difficult to
reproduce bugs would benefit from a log being generated by default as it
may be difficult to come by after the fact.

Log to ${TMPDIR}/cooker.log.${DATETIME} by default.

[YOCTO #1771]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agobase.bbclass: Use bb.plain to print the build header
Richard Purdie [Thu, 24 Nov 2011 14:48:58 +0000 (14:48 +0000)]
base.bbclass: Use bb.plain to print the build header

If we just use print, the bitbake logging functions don't see the message
and it can get lost if bitbake is for example logging messages to disk.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agolibtool: Avoid relinking when cross compiling, its pointless
Richard Purdie [Wed, 16 Nov 2011 16:58:34 +0000 (16:58 +0000)]
libtool: Avoid relinking when cross compiling, its pointless

There is no point in having "executable" binaries in the .libs
directory linked with different rpaths to the target which
could concivably be run on the build system when cross compiling.

This patch removes the extra rpaths ($compile_rpath) so that the
output from the "link" stage can be used on the target. We can then
avoid having to "relink" during the install stage.

This saves some build time (do_install is over 2 minutes faster for
pulseaudio).

This patch also removes an annoying "seems to be moved" warning
which is totally bogus in the sysroot case.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agoapt: Fix locale header and hardcoded libname issues
Richard Purdie [Wed, 23 Nov 2011 23:24:20 +0000 (23:24 +0000)]
apt: Fix locale header and hardcoded libname issues

apt wasn't building on modern libc/compiler combinations due to missing
header includes.

The libcpp version was also being hardcoded, this patch generates it
dynamically to work on different host systems which no longer have
this.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agodirectfb: change check sum
Xiaofeng Yan [Wed, 23 Nov 2011 07:05:24 +0000 (15:05 +0800)]
directfb: change check sum

Change check sum of directfb. The previous checksum was incorrect when it
was added due to a corrupted download on the submitter's system.

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agodistro_tracking_fields: Update after Manual Check
Saul Wold [Tue, 22 Nov 2011 21:18:54 +0000 (13:18 -0800)]
distro_tracking_fields: Update after Manual Check

Checked upstream for updates of the following packages:
 - libiconv
 - libxml2
 - sysstat
 - tcp_wrapper
 - texinfo

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agodistro_tracking: update manual checks for recipes
Shane Wang [Tue, 22 Nov 2011 12:45:36 +0000 (20:45 +0800)]
distro_tracking: update manual checks for recipes

Updating manual checks for jpeg, tinylogin, eggdbus, setserial, acpid, menu-cache, x11vnc, etc.

Signed-off-by: Shane Wang <shane.wang@intel.com>
13 years agodistro_tracking: update manual checks for recipes
Shane Wang [Tue, 22 Nov 2011 07:48:21 +0000 (15:48 +0800)]
distro_tracking: update manual checks for recipes

Updating manual checks for minicom, libtirpc, opkg, which, quota, rpcbind, strace, wireless-tool, irda-utils, libical, flac, libomxil, libsamplerate0, etc.

Signed-off-by: Shane Wang <shane.wang@intel.com>
13 years agodistro_tracking: update manual checks for recipes
Shane Wang [Fri, 18 Nov 2011 07:23:57 +0000 (15:23 +0800)]
distro_tracking: update manual checks for recipes

Updating manual checks for hdparm, screen, watchdog, freetype, oprofile, and tcf-agent.

Signed-off-by: Shane Wang <shane.wang@intel.com>
13 years agoat: update the upstream-status of patches
Nitin A Kamble [Thu, 17 Nov 2011 20:01:30 +0000 (12:01 -0800)]
at: update the upstream-status of patches

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
13 years agodistro_tracking: update manual checks for various recipes
Scott Garman [Mon, 14 Nov 2011 23:37:18 +0000 (15:37 -0800)]
distro_tracking: update manual checks for various recipes

Updating manual check dates for the following recipes: tar, libpng,
expat, dosfstools, e2fsprogs

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
13 years agodistro_tracking: Refect Recipe Updates & Status
Saul Wold [Wed, 9 Nov 2011 22:14:01 +0000 (14:14 -0800)]
distro_tracking: Refect Recipe Updates & Status

 * libnl    - NO_UPDATE_REASON due to incompatibility
 * zlib     - has wrong version in update list (121)
 * libtasn1 - Update to 2.10
 * pkgconfig - NO_UPDATE_REASON due to removal of glib-conf
 * file     - update to 5.09
 * dchp     - New version is 4.2.3, not updated yet.
 * tiff     - NO_UPDATE_REASON wait until 4.0.0
 * gobject-interopsectio - NO_UPDATE_REASON can not cross-build
 * gnu-config  - Udpate to git HEAD - requires ASSUME_PROVIDED += "git-native"

Signed-off-by: Saul Wold <sgw@linux.intel.com>
13 years agogconf: Add note to header about patch filed upstream
Richard Purdie [Wed, 23 Nov 2011 15:46:18 +0000 (15:46 +0000)]
gconf: Add note to header about patch filed upstream

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agogconf: Ensure the correct backend directory is used
Richard Purdie [Wed, 23 Nov 2011 15:05:39 +0000 (15:05 +0000)]
gconf: Ensure the correct backend directory is used

Without these changes, gconf will use the hardcoded backend directory meaning
we can see errors when building if the binary was relocated.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agogconf: Ensure that GCONF_SCHEMA_INSTALL_SOURCE uses correct paths
Richard Purdie [Wed, 23 Nov 2011 08:48:31 +0000 (08:48 +0000)]
gconf: Ensure that GCONF_SCHEMA_INSTALL_SOURCE uses correct paths

The gconf .m4 files use the path encoded into the gconf-native tools
by default to populate this variable. This doesn't work if we're
relocating the gconf-native binaries. By setting this variable we
ensure that the m4 files always pick up the correct path.

This fixes errors seen on the yocto autobuilder.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agomatchbox-stroke: Update to latest git revision to add missing Xrender dependency
Richard Purdie [Wed, 23 Nov 2011 08:47:07 +0000 (08:47 +0000)]
matchbox-stroke: Update to latest git revision to add missing Xrender dependency

Without this, linking can fail due to missing libraries since the
code depends on Xrender. The fix was already upstream, we just need
to use a revision that includes it.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agomatchbox-wm: Fix variable type in _NET_WORKAREA setting
Zhai Edwin [Tue, 22 Nov 2011 06:15:08 +0000 (14:15 +0800)]
matchbox-wm: Fix variable type in _NET_WORKAREA setting

According to XChangeProperty doc, array of "long" should be used when format is
32. Wrong _NET_WORKAREA parameter caused blank screen in matchbox-desktop on 64
bit platform.

[YOCTO #1689] got fixed.

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agostaging: Remove debug accidently left in
Richard Purdie [Tue, 22 Nov 2011 15:00:46 +0000 (15:00 +0000)]
staging: Remove debug accidently left in

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agostaging.bbclass: Add BB_SETSCENE_VERIFY_FUNCTION function
Richard Purdie [Mon, 21 Nov 2011 14:37:10 +0000 (14:37 +0000)]
staging.bbclass: Add BB_SETSCENE_VERIFY_FUNCTION function

Since we clean out do_populate_sysroot if do_configure runs, don't
allow do_populate_sysroot_setscene functions if we're going to
run do_configure.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agopackage.bbclass: Ensure paths to rpmmarcos and rpmpopt are set
Richard Purdie [Tue, 22 Nov 2011 13:01:42 +0000 (13:01 +0000)]
package.bbclass: Ensure paths to rpmmarcos and rpmpopt are set

If rpm-native was built in an alternative location, it may not relocate correctly
unless the rpmpopt and macros paths are explicitly specified.

This fixes errors seen on the Yocto autobuilder where pkgconfig
"provides" entries could disappear leading to image dependency failures.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agogcc-4.6: Fix gcc ICE on qt4-x11-free/armv7-a
Khem Raj [Sun, 20 Nov 2011 17:50:47 +0000 (09:50 -0800)]
gcc-4.6: Fix gcc ICE on qt4-x11-free/armv7-a

Backport fix for PR 47551 fixes the ICE seen on armv7-a/qt4-x11-free
Bump up SRCREV past gcc 4.6.2 release

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agodbus: update to 1.4.16
Koen Kooi [Sun, 20 Nov 2011 09:52:45 +0000 (10:52 +0100)]
dbus: update to 1.4.16

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agomime.bbclass: Add --disable-update-mimedb to EXTRA_OECONF
Andreas Müller [Sun, 20 Nov 2011 21:09:23 +0000 (22:09 +0100)]
mime.bbclass: Add --disable-update-mimedb to EXTRA_OECONF

* ensure update-mime-database is not called during build time
* this patch was included in meta-oe [1] about the same time when mime.bbclass
  was migrated from meta-oe -> oe-core so it seems it got lost.
* tests: build from scratch / run on overo / additional check: no unpacked
  files for gnome-control-center

[1] http://cgit.openembedded.org/meta-openembedded/commit/meta-oe/classes?id=6b765989a42ab314d1611f4dec78b07b562a9e7d

Signed-off-by: Andreas Müller <schnitzeltony@gmx.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agopulseaudio-0.9.23: Fix build on uclibc
Khem Raj [Sun, 20 Nov 2011 17:50:45 +0000 (09:50 -0800)]
pulseaudio-0.9.23: Fix build on uclibc

Import patch from oe.dev rev 7c33f2e906a20e139d53b4f2d8fbc2773a4725b3
to fix bluez build when using uclibc

cgit link here

http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=7c33f2e906a20e139d53b4f2d8fbc2773a4725b3

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agonetbase: Update recipe 4.46 -> 4.47
Khem Raj [Sun, 20 Nov 2011 17:50:46 +0000 (09:50 -0800)]
netbase: Update recipe 4.46 -> 4.47

moreover 4.46 tar is not available on debian pools

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agogettext_0.18.1.1: Fix QA Issue
Khem Raj [Sun, 20 Nov 2011 17:50:44 +0000 (09:50 -0800)]
gettext_0.18.1.1: Fix QA Issue

For uclibc builds we get the following QA error

ERROR: QA Issue: non -dev/-dbg/-nativesdk package contains symlink .so:
gettext-runtime path
'/work/armv5te-oe-linux-uclibceabi/gettext-0.18.1.1-r4/packages-split/gettext-runtime/usr/lib/libintl.so'

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agopackage.bbclass: fix path for relative links on elf files, when moved to debugdir
Henning Heinold [Sun, 20 Nov 2011 21:32:53 +0000 (22:32 +0100)]
package.bbclass: fix path for relative links on elf files, when moved to debugdir

* relative links for elf files like ../foo.so ends up in the
  debugdir with ../.debug/foo.so, this causes infinite fileaccessloops
  fix it by adding an extra "../" to the link path

Signed-off-by: Henning Heinold <heinold@inf.fu-berlin.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agosstate.bbclass: Ensure we expand stamp-extra-info
Richard Purdie [Mon, 21 Nov 2011 14:00:05 +0000 (14:00 +0000)]
sstate.bbclass: Ensure we expand stamp-extra-info

Without this change we can end up looking for <stamp>.${MACHINE}
instead of the expected expanded value.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agolibtool-cross: Unbreak and actually use more of it
Richard Purdie [Wed, 16 Nov 2011 23:59:42 +0000 (23:59 +0000)]
libtool-cross: Unbreak and actually use more of it

We should be using libtool-cross for cross compiling but
were not. This patch sets datadir so libtoolize ends up
containing correct paths. It then installs libtoolize.

The path ltmain.sh was installed to was incorrect and this is fixed.

We also now install all the libtool m4 macros and config files.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agouseradd: Ensure -native recipes don't depend on target recipes
Richard Purdie [Mon, 21 Nov 2011 13:59:19 +0000 (13:59 +0000)]
useradd: Ensure -native recipes don't depend on target recipes

Without this change, dbus-native can end up depending upon base-passwd
for example. This change mirrors the existing nativesdk code.

Based on a patch from Henning Heinold <heinold@inf.fu-berlin.de>
but with some additions from me.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agomesa: fix calling host's commands
Kang Kai [Fri, 18 Nov 2011 06:36:39 +0000 (14:36 +0800)]
mesa: fix calling host's commands

mesa use its script file bin/mklib to generate libraries, and call the
host's commands "ar/ranlib/gcc/g++". Fix it to call the cross-compile
tools properly.

Signed-off-by: Kang Kai <kai.kang@windriver.com>
13 years agoshared-mime-info: fix ordering of PACKAGES
Koen Kooi [Fri, 18 Nov 2011 14:13:33 +0000 (15:13 +0100)]
shared-mime-info: fix ordering of PACKAGES

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
13 years agobootimg: Create a valid boot sector for the iso image
Damien Lespiau [Thu, 10 Nov 2011 18:29:50 +0000 (18:29 +0000)]
bootimg: Create a valid boot sector for the iso image

It's possible to create an "hybrid" iso image that you can both burn
to a CD-Rom and dd to a USB key. isohybrid will create a valid boot
sector for the USB key case.

[YOCTO #1763]

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agoglib-2.0: fix qsort_t testing for uclibc
Henning Heinold [Fri, 18 Nov 2011 11:59:48 +0000 (12:59 +0100)]
glib-2.0: fix qsort_t testing for uclibc

* instead of disabling the configure test and set it always true
  use the site files to set it true for glibc/eglibc and false
  for uclibc
* remove the patch
' tested only for uclibc
* bump PR

Signed-off-by: Henning Heinold <heinold@inf.fu-berlin.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agosanitytest: remove rpm/zypper tests if PACKAGE_CLASSES does not set package_rpm
Jiajun Xu [Thu, 17 Nov 2011 06:05:50 +0000 (14:05 +0800)]
sanitytest: remove rpm/zypper tests if PACKAGE_CLASSES does not set package_rpm

If PACKAGE_CLASSES does not set package_rpm as the first item, the root filesystem
will not be generated based on rpm. We need remove rpm/zypper tests against
non-rpm filesystem.

[YOCTO #1757]

Signed-off-by: Jiajun Xu <jiajun.xu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agosanitytest: use different option for command ps from busybox and procps
Jiajun Xu [Thu, 17 Nov 2011 06:05:49 +0000 (14:05 +0800)]
sanitytest: use different option for command ps from busybox and procps

Current sanitytest use option -e for ps command, which only works for ps
from procps. It fails if ps is provided by busybox. Add check to use different
option for command from busybox and procps.

[YOCTO #1756]

Signed-off-by: Jiajun Xu <jiajun.xu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 years agopackage_rpm: Set _tmppath to avoid races over tmp files
Richard Purdie [Wed, 16 Nov 2011 16:59:29 +0000 (16:59 +0000)]
package_rpm: Set _tmppath to avoid races over tmp files

Occasionally we keep seeing "unable to open temp file" messages during
do_package_write_rpm tasks. This appears to happen when multiple
processes are writing rpm files and is likely due to using the
shared system temp directory. This patch changes the tmp path
to the package work directory meaning conflicts should become
a non-issue.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>