bash: fix mkbuiltins build failure

On hosts with FORTIFY_SOURCES, stringize support is required, as it's used by
the macros to wrap functions (e.g. read and open in unistd.h). Those wrappers
use the STRING() macro from unistd.h. A header in the bash sources overrides
the unistd.h macro to 'x' when HAVE_STRINGIZE is not defined, causing the
wrappers to generate calls to 'xread' and 'xopen', which do not exist,
resulting in a failure to link.

Assume we have stringize support when cross-compiling, which works around the
issue.

It may be best for upstream to either give up on supporting compilers without
stringize support, or to not define STRING() at all when FORTIFY_SOURCES is
defined, letting the unistd.h one be used, instead.

(From OE-Core rev: f7a25dd72d1d463eb72d48c6f9dd968d376496c0)

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sstate: Fix various path manipulation issues

Fix missing parameter to endswith and pass paths through normpath to remove
any duplicate "/" characters which would corrupt other calls like basename.

(From OE-Core rev: 172a74c540378149eec493c37c030e9f42f9603d)

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

bitbake: knotty: kill duplicated import of 'time'

The duplicated import could result in an UnboundLocalError.

(Bitbake rev: a098cebd5c33ebd704efd35d9e655262283cbe1f)

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

utils.bbclass: Fix documentation of create_cmdline_wrapper

(From OE-Core rev: 56160ca49dd546b7db07ae2021eefef7279b0f10)

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

classes: Be consistent about sstate-inputdirs/outputdirs ending with '/'

If sstate-inputdirs and sstate-outputdirs don't match with ending '/'
characters, the manifest file can end up corrupted. This change
ensures the metadata is consistent in ending do_populate_root tasks
with this character to avoid manifest file corruption.

(From OE-Core rev: 3910eaf88d14904eef85b9e391387547df7fc54e)

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

sstate: Be consistent about sstate-inputdirs/outputdirs ending with '/'

The manifest file can become corrupted if sstate-inputdirs and sstate-outputdirs
don't have matching endings. This patch ensures that even if set incorrectly,
the code functions as intended, thereby handling manifest corruption safely.

(From OE-Core rev: 0109a3623a19f9ae289952a4f054e53c3eca4eaa)

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

python.inc: make lsb override more concise

No functionality change, just a readability improvement

(From OE-Core rev: 7a27f95c2800285d7f97fead616620bfd7dabbe3)

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

libc-common: Ensure sysconfdir exists before installing files to it

Depending on the eglibc configuaration, the directory may or may not exist.

(From OE-Core rev: aa89b80a42297196c7bbba55fe2396ba1f98acc7)

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

boost: Activate zlib and bzip2 because they now work.

This patch is needed for making wesnoth(will be sumbited
  in meta-games) work.

[RP: Add missing bzip2 dependency]
(From OE-Core rev: 9987998d7431fb33b12a7e210024a8193b43e717)

Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@no-log.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: print clear message for "bitbake -e ASSUME_PROVIDED"

"bitbake -e ASSUME_PROVIDED" should fail, but the error message wasn't
clear enough in the past:

$ bitbake -e bzip2-native
[snip]
ERROR: Command execution failed: Traceback (most recent call last):
  File "/buildarea/lyang1/poky/bitbake/lib/bb/command.py", line 94, in
runAsyncCommand
    commandmethod(self.cmds_async, self, options)
  File "/buildarea/lyang1/poky/bitbake/lib/bb/command.py", line 323, in
showEnvironmentTarget
    command.cooker.showEnvironment(None, pkg)
  File "/buildarea/lyang1/poky/bitbake/lib/bb/cooker.py", line 325, in
showEnvironment
    fnid = taskdata.build_targets[targetid][0]
KeyError: 0
[snip]

With this patch, the massage will be:
[snip]
ERROR: bzip2-native is in ASSUME_PROVIDED
ERROR: Command execution failed: Exited with 1

Summary: There were 2 ERROR messages shown, returning a non-zero exit code.

[YOCTO #3392]

(Bitbake rev: f31447dac92454c822d4ebb7dd48e96c6c69dde4)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: hob was freezing because it doesn't receives well the log file

-after pressing "build image" button, hob was freezing because it didn't
receive well the log file

[YOCTO #3398]
(Bitbake rev: e3619e34d43c3f7725fc83c362d8cbd07e153ebe)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mesa-demos: fix --with-glut check

The GLUT check was automatic and couldn't be disabled, so mesa-demos would gain
a GLUT dependency if it was present when built.

So, fix configure.ac so that --without-glut works as expected.

(From OE-Core rev: fa7fb44d1ca2b8a57509806bde19672c68ef157d)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

initrdscripts: fix udevd in the live boot init scripts

udevd moved location and isn't in $PATH anymore, so use an absolute path to
start it.

The control socket path moved too, so mkdir the directory it's in.

Mounts the new devtmpfs on /dev device tree.

(From OE-Core rev: 543606e5dc379497c34196d004a214e847f5db2d)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexandru Damian <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libunique: fix compilation with GLib 2.34

The new GLib has deprecated some functions, and libunique was building with
-Werror.  Take a patch from upstream to update the build system and rationalise
the warning flags.

(From OE-Core rev: 713e1c82d32a477be84951d1dba8326b0342c8ca)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

insane: detect and warn about relocations in .text

(From OE-Core rev: bc08838ddab0d16d889f81244529a0302a9240bc)

Signed-off-by: Phil Blundell <pb@pbcl.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libxslt: upgrade to 1.1.27

pkgconfig_fix.patch: adapted to the new version

(From OE-Core rev: 960ad0998a9edbefe57d08b6f587682f0dc9d768)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

puzzles: upgrade to r9660

License checksum change due to diff:
4,5c4,5
< K�lker, Dariusz Olszewski, Michael Schierl, Lambros Lambrou and
< Bernd Schmidt.
> K�lker, Dariusz Olszewski, Michael Schierl, Lambros Lambrou, Bernd
> Schmidt and Steffen Bauer.

(From OE-Core rev: dbf89870ae2eade8210ef6bbc9c13be17336ea68)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mesa: remove python2 detection hack

Our python-native is 2.7.3 which ships python2, and we've been patching it in to
earlier versions since September 2011.

(From OE-Core rev: 1961ea1493caf5fe2959db2a11831a4a5a17eaeb)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk_base.bbclass: use SDK_ARCH instead of SDKMACHINE

If SDKMACHINE is not set, the toolchain will be built but the tarball
installer will not run. A better choice is to use SDK_ARCH because, even
if SDKMACHINE is not set, SDK_ARCH is set, by default, to BUILD_HOST.

(From OE-Core rev: b6d391903ae8baf902fa142a58533857ade6afd3)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glib-2.0: upgrade to latest stable, 2.34.1.

Also explicitly disable the test suite (as we can't run it), subsequently
dropping 60_wait-longer-for-threads-to-die.patch and nodbus.patch.

nolibelf.patch has been merged upstream, drop.

Upstream has dropped the pre-generated man pages, to generate them again we'd
need libxslt and the DocBook infrastructure.  We can live without the man pages
as those build-dependencies are non-trivial.

(From OE-Core rev: ce5fcad59fff19dbffc2d7b49c0c8bf3701d17ed)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glib-2.0: remove zip build dependency

This dependency mysteriously appeared a long time ago for no good reason, remove
it.

(From OE-Core rev: 5e2fdfcbe71f0b6a786ef609e288c0999000e163)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt4: remove -lGLU from QMAKE_LIBS_OPENGL in our linux.conf

upstream does not need GLU since:
commit e7eed096a0c33607a7a37baaf06e5952dc9d556b
Author: Bj<C3><B8>rn Erik Nilsen <bjorn.nilsen@nokia.com>
Date:   Mon Aug 9 14:07:01 2010 +0200

    Remove dependency of OpenGL Utility Library (GLU).

    GLU is not part of standard OpenGL and is not used internally in Qt,
    so we should not depend on it.

    Task-number: QT-12227
Reviewed-by: kim
(From OE-Core rev: 181874ba8033535bb4ce2b36725ae0a71c27b3bd)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xorg-driver-common: remove AC_CHECK_FILE workaround

The drivers don't generally use AC_CHECK_FILE anymore, so remove this
brute-force kludge. Any subsequent breakage can be worked-around in the recipe
and fixes submitted upstream.

(From OE-Core rev: 1b0d9cb1801a8eb68c82dfcda5a1da420ac8dd83)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

meta: remove redundant _FOR_BUILD variables

(From OE-Core rev: acabd2158d9004dedfdfad8c170b77d32684f3fc)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

autotools: set _FOR_BUILD variables here

(From OE-Core rev: edf30561184ec42e5692a55fdf93304fac0fdb1b)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/classextend: Ensure we don't extend expressions more than once

We could end up with MLPREFIX being prepended to variables like
PACKAGE_DYNAMIC. This patch avoids the problem and unbreaks builds.

[YOCTO #3389]

(From OE-Core rev: 5e4714a454f9f742bf8af70dad2aa66ccb87e3fd)

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

gnome-desktop: Now we depend on gnome-common-native, use the correct sysroot

This fixes the build after gnomebase was changed to depend on
gnome-common-native.

(From OE-Core rev: e57278dfde1f975fa34e72354749bbc91380f36b)

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

gnomebase: depend on gnome-common-native

gnome-common is a build-only dependency so we should depend on the native
variant.  This also resolves an (incorrect) GPLv3 license issue in gnome-common
at build-time.

This will also remove the pointless gnome-common-dev RRECOMMENDS in any -dev
package that uses gnomebase.

(From OE-Core rev: 6a4f394bc1280f5d58d928a2f7cff7cce4eb3b2b)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk_base.bbclass: check installation machine before installing SDK

Do not allow installer to continue if the installation machine architecture
does not match the intended SDK machine architecture.

[YOCTO: #3269]

(From OE-Core rev: 1f78e2c97f978f0f02e884870e7c495751f0802c)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnome-common: Fix license

gnome-common 2.28 is GPLv2+.  From Christian Persch, upstream:

  The licence is presumed GPL2+, although it's not there explicitly.  GPL2+
  because as far as I could figure out when I tried to, gnome-autogen started in
  gnome-core which had a GPL2 COPYING file.

(From OE-Core rev: 1d1532f495041ac58aeafd06781ac87ee3bd3f6a)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake.conf: Add udev rules into ${PN} files by default

As we move to systemd, udev is not provided by systemd where the arch
independent files are stored in /lib and /usr/lib and not in
${base_libdir} and ${libdir} which means the files like udev rules
go into /lib/udev or /usr/lib/udev. This patch adds these paths
to be packaged into default ${PN} output package from a recipe

(From OE-Core rev: 6cbc0c7bf3f35e18f0abd29e5a704fba55f88ab2)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

create-recipe: update re pattern and output

In the URL, there may be more than just digits in the version section,
something like xz 5.1.2alpha. Update RE pattern to catch all the string
after package name and before '.tar' in URL as package version.

And error message which has been sent to /dev/null still shows on Ubuntu
12.10 with perl 5.14. Update the way to find source tar file to eraser
the error message.

configure files may rewrite the version section, and that is not necessary.
Test when version section has been set, omit the version value from
configure files.

And tweak for output to bb file.

(From OE-Core rev: 17f09ab713acc814ec0561b1c41fa87d9bf7b83f)

Signed-off-by: Kang Kai <kai.kang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

hwclock.sh: improve hwclock.sh script to use UTC variable

Make UTC variable in /etc/default/rcS has effect on hwclock.sh.
This variable declares whether the Hardware Clock is kept in UTC
or local time. Default its value to "yes" and change the comment.

[YOCTO #3341]

(From OE-Core rev: 2bda4bbda3f3033cfb324778ef88f2aedad4a83b)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

recipes-devtools: replace virtclass-native(sdk) with class-native(sdk)

The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.

[YOCTO #3297]

(From OE-Core rev: bb67ddeb2eed3e25c626a279ef53a7e8c7bfe6f2)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

recipes-extended: replace virtclass-native(sdk) with class-native(sdk)

The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.

[YOCTO #3297]

(From OE-Core rev: 528b4ab831c7b0bc1412318d29e2b7f9cf711d57)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

recipes-graphics: replace virtclass-native(sdk) with class-native(sdk)

The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.

[YOCTO #3297]

(From OE-Core rev: db1a03da3a6a6e7adb68e28883204adfaa8b3f47)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

recipes-support: replace virtclass-native(sdk) with class-native(sdk)

The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.

NOTE:
There were 2 errors in libcap.inc, the BUILD_LDFLAGS_virtclass_native
should be BUILD_LDFLAGS_virtclass-native (the "_" should be "-"),
otherwise it doesn't work, and the value was: "-Wl,rpath=...", this is
incorrect, it shoudl be: "-Wl,-rpath=..." (lacked a - ), but we don't
need this line, since it is already in the default BUILD_LDFLAGS. Remove
it and we don't need to bump the PR since we just removed a unused line.

[YOCTO #3297]

(From OE-Core rev: cafb550fe9034754933f1708446dde155dcc3d51)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

recipes-gnome: replace virtclass-native(sdk) with class-native(sdk)

The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.

[YOCTO #3297]

(From OE-Core rev: d37944fff256e9f52d05a56db3edb42c7a352cce)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

recipes-kernel: replace virtclass-native(sdk) with class-native(sdk)

The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.

[YOCTO #3297]

(From OE-Core rev: 7c7da0af592647be0ab9ab37a8ad2a7ce4890a46)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

recipes-connectivity: replace virtclass-native(sdk) with class-native(sdk)

The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.

[YOCTO #3297]

(From OE-Core rev: 37429a94133c0d0bfae71d1d4329aee6dd5eb98b)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes: replace virtclass-native(sdk) with class-native(sdk)

The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.

[YOCTO #3297]

(From OE-Core rev: 9fbeab63315fef0dbcc91c5e7051665764758a6e)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

conf: replace virtclass-native(sdk) with class-native(sdk)

The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.

[YOCTO #3297]

(From OE-Core rev: 651c6fd6ef8be17ecba53f8054d24d1077198d5b)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

recipes-core: replace virtclass-native(sdk) with class-native(sdk)

The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.

[YOCTO #3297]

(From OE-Core rev: 5193485a42dfb3396d0f12aaa7732c5db29d7338)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

babeltrace: fix depends

(From OE-Core rev: 0c1c0a890f6e99608eda3d20a6b89027a2c72e96)

Signed-off-by: Jack Mitchell <jack.mitchell@dbbroadcast.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-core-tools-profile: include sysprof only if x11 is distro feature

(From OE-Core rev: 9dd277e3276aef918e4bb84322043db0b4e68e11)

Signed-off-by: Jack Mitchell <jack.mitchell@dbbroadcast.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

latencytop: disable GTK and remove gtk+ dependacy

When DISTRO_FEATURES does not contain x11, disable GTK GUI and
also don't add gtk+ to the DEPENDS

(From OE-Core rev: 1b41a4660d0dd8f1863bb94d08562927cc108ee1)

Signed-off-by: Jack Mitchell <jack.mitchell@dbbroadcast.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mesa: Build separate GLU library

Mesa has removed GLU from the core tree upstream, so remove it from the
Mesa build and add the separate tarball as a new recipe.

(From OE-Core rev: 4395691a44b198ba0b9a969cbade669e8de07a4f)

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

terminal: Ensure existing environment exports are preserved in devshell

After recent changes to terminal.bbclass, variables like PATH were no longer
preserved within the devshell. This change ensures they are inherited into
the environment of devshell and PATH for example has the correct values.

(From OE-Core rev: f2dfc50bdf403719d40d04488245fd37655b5480)

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

terminal.bbclass: Ensure parent environment is set

If this isn't done, various terminals fail to launch correctly
with "No such file or directory" errors. This adds back the environment
manipulation removed in the addition of "custom" terminal command
support but shouldn't regress that additional functionality

(From OE-Core rev: 424d2339b462081010af6e7525a71f64d97ff05e)

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

bitbake: command: add error to return of runCommand

Currently, command.py can return an error message from runCommand, due to
being unable to run the command, yet few of our UIs (just hob) can handle it
today. This can result in seeing a TypeError with traceback in certain rare
circumstances.

To resolve this, we need a clean way to get errors back from runCommand,
without having to isinstance() the return value. This implements such a thing
by making runCommand also return an error (or None if no error occurred).

As runCommand now has a method of returning errors, we can also alter the
getCmdLineAction bits such that the returned value is just the action, not an
additional message. If a sync command wants to return an error, it raises
CommandError(message), and the message will be passed to the caller
appropriately.

Example Usage:

    result, error = server.runCommand(...)
    if error:
        log.error('Unable to run command: %s' % error)
        return 1

(Bitbake rev: 717831b8315cb3904d9b590e633000bc897e8fb6)

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: providing details about process state through porgress bar

-small changes to the text of the progress bar when parsing recipes

[YOCTO #3282]
(Bitbake rev: 90c0dfc39c3ce13e53c7c91168dc3401f7df476b)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake/server: Remove dead console log code

This code is dead and doesn't do anything so lets remove it.

(Bitbake rev: 8d45739f49618757a5d7d79782deda355e3981ec)

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

bitbake: cooker.py: Don't dump the environment into the console log file

Dumping the environment data into the console log files directory is
invariably not what the user wants or expects and leads to confusion
when looking at the log directory.

This change forces the logs to be disabled by default when using
the -e option.

(Bitbake rev: 5d825b31d1133e41d3982db1b94f6a30a6fb99f7)

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

bitbake: uihelper: Set update flag when start event encountered

Its a minor correctness detail but the update flag should be set
when Start events are encountered.

(Bitbake rev: 96683ed68e11672ff22fb4a291d2628676c136f0)

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

bitbake: knotty: Ensure last tasks are disaplyed correctly in the footer

There was an issue where the last tasks run by bitbake would not
correctly get displayed in the knotty footer. This was due to the
total count including active tasks. This change ensures the footer
is displayed if the are any running tasks.

(Bitbake rev: d787e4efc106589811651bc18ca48d5223443b95)

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

openssl: Use ${CFLAGS} not ${FULL_OPTIMIZATION}

The latter variable is only applicable for target builds and could
result in passing incompatible options (and/or failing to pass
required options) to ${BUILD_CC} for a virtclass-native build.

(From OE-Core rev: 0e90a303bc5cb0ede21ff4346843f9daeddfff45)

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev: Ensure tmpfs are mounted and volatile/run exists

There is a race with udev where eiher the run directory can get replaced
during bootup leading to ude errors, or if the tmpfs was mounted and
populate-volatiles hasn't run, udev won't start at all.

This ensures that any tmpfs get mounted before udev starts and that the
default volatiles/run directory at least exists, fixing the races
and boot time errors caused after the recent udev upgrade.

(From OE-Core rev: fbec192f6bc41a335ede85843ba22a89d13501ab)

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

udev: Use correct variable in udev.conf for run_path

(From OE-Core rev: 6e4a1743a88fe6a002ca8f784b2dad3f493984b5)

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

polkit: remove license.html from LIC_FILES_CHKSUM

This is a generated file and gets removed by "make clean" which then
causes subsequent rebuilds to fail.  Also, the content in this file
is taken verbatim from COPYING (which is already in LIC_FILES_CHKSUM)
so checking it for a second time doesn't accomplish much.

(From OE-Core rev: df7817649cc62bfef04a5c1aab97f7964e6bd76c)

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: reordering the layers in the Hob Layers dialog

-since the order of the layers can potentially impact
the build outcome, users should be able to reorder
the layers within the layers dialog;
-used TreeView Drag and Drop

[YOCTO #3270]
(Bitbake rev: 2bd9a00facb90f7d76a9bdaa86ca765fb2159e71)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob/hig: Hob doesn't save properly proxy settings

User introduced proxy settings were not saved
when a proxy details input dialog was opened.
The proxy settings were lost upon return, and
restored from the previously stored one.

Also:
Code cleanup:
details_cb() function duplicate definition
removed

Fixes [YOCTO #3240]

(Bitbake rev: aa64b2e472f8a9713e3dc25647aa2cae474e2622)

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: during recipe parsing, progress bar text provides details about the process state

- indicating on progress bar that hob has gone from parsing
recipes to "Generating dependency tree"; this will provide
some visibility of what has caused the "stop" button state
to change

[YOCTO #3282]
(Bitbake rev: d964d04ff1f59a590bd3ab5430517d79e92536d0)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: add a progress indicator when you select 'view log'

- created a new file named "hobthreads.py", defining a thread
for opening the log file in a subprocess using subprocess module;
in the future I think we will add some other threads here, to
implement some other performance issues
- on "builddetailspage", "packageselectionpage" and "imagedetailspage"
I have changed the manner for opening the log file; it uses the thread
to open the file, and on main thread it creates a dialog to show a
progress bar, which pulses till the file is open
- this was added because when the log file is big, it takes time to
be opened; on the dialog you can use "Cancel" button to terminate the
process initiated to open the file

[YOCTO #2997]
(Bitbake rev: 165362a63f085991b6bab63ab90a0c7b9bf6b784)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev-extraconf: Don't mount root filesystem under /media

The mount.sh handler attempts to prevent already-mounted filesystems
from being mounted as dynamic/removable "/media".  But it misses the
case where the kernel has mounted the root filesystem (e.g. with
"root=/dev/sda1").  In that situation, /proc/mounts has a device name
of "/dev/root" instead of the proper $DEVNAME string exposed by udev.
So we must also test the root filesystem device number vs. the
$MAJOR/$MINOR udev tells us.

(From OE-Core rev: 3543d0db691e82098c1da7bf12f43e0c57551a3d)

Signed-off-by: Andy Ross <andy.ross@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

busybox: add /usr/bin/stat applet

The busybox defconfig lacks a stat tool, the functionality of which
cannot be reproduced in a way accessible to a shell script running in
a minimal configuration.  Enable, and modify the installation path to
/usr/bin/stat to match the coreutils tool for proper alternatives
handling.

(From OE-Core rev: ef7e1239b95dbef4e461007d6d0612c27a7919ec)

Signed-off-by: Andy Ross <andy.ross@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lttng-modules: remove unused lttng-syscalls patch

commit b7e184508 (lttng-2.0: fix srcrev/pv to match the recipe
filenames) removed the
lttng-sycalls-protect-is_compat_task-from-redefiniti.patch from the
SRC_URI but forgot to remove the patch itself.

(From OE-Core rev: 6745b927a40e523cfda4ce2ca6422d69a6791e8a)

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake.conf: Change systemd_unitdir definition

systemd_unitdir indicates the arch independent
files which are basically scripts and unit files
and systemd wants then to be in /lib always even
when base_libdir is  /lib64, hence we have to reflect
that and not use base_libdir to define it. Otherwise
on architectures where base_libdir is lib64 e.g. ppc64
or multilibbed x86_64 this wont work

(From OE-Core rev: 50e713d4ae35f9b5f5f2a515a95d77573610d707)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

distro_identifier: replace slash with hyphen

Use "-" instead of "/" in "n/a" strings ("Distributor ID" and/or
"Release"), provided by `lsb_release`.
This leads to directories and subdirectories created in ./sstate-cache/
e.g. Distro-n/a/ where "Distro-n" is dir and "a" is subdir.

(From OE-Core rev: c2a9f0c330f0da82792f87e3860b06c25b986983)

Signed-off-by: Mihai Lindner <mihaix.lindner@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk_base: allow SDK path of various level

In the previous version, tar extraction use the --strip-component
option with "4" hard coded value.
If we set another SDKPATH, with a different depth, the sdk installation
fails.

This patch computes the level from the SDKPATH value.

(From OE-Core rev: 7aee4e9438755c230e1399bd5226d6c8e7fcca31)

Signed-off-by: Fabien Proriol <fabien.proriol@jdsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

ltp: add perl to RDEPENDS

Some of the ltp scripts are perl, this was not seen in the
past because ltp is normally installed in an -sdk build with
perl already there.

(From OE-Core rev: 930216cb9092904642c0419d3475fc731ab0694b)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sstate: add manifest info for shared file matches

Present the manifest file that contains the matches for
files being installed to a location that already contains
that file. This will help to determine which is the correct
recipe to fix when this occurs.

[YOCTO #3191]

(From OE-Core rev: 56268f6e4ed1fc11143173bb1717a8be78c728a5)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gst-meta-base: include ximagesink plugin when x11 is a feature

(From OE-Core rev: 6e44bf71fd4645ea09b3e34b907b17e92b5225e3)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio: remove libpulse-browse, removed upstream in 2011

(From OE-Core rev: 24bf52ec5e53339f15573e9c9789b7ce14f743bd)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio: fix library packaging

libpulsecommon has moved into libdir/pulse.

libpulsedsp is an internal library for padsp so put that into pulseaudio-misc
along with padsp.

(From OE-Core rev: 01b9b1edcc82e0c8cb2240359daa5b14d2c0146b)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio: move helper binaries into the relevant module packages

proximity-helper is only used by the bluetooth-proximity module, and
gconf-helper is only used by the gconf module.  Clarify the packaging and clean
up dependencies by shipping the helper binaries with the modules that spawn
them.

(From OE-Core rev: 039170824cb77c1a68ec91d9f4dc1ae12f701b87)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio: add GConf dependency

Add explicit build-dependency on GConf as currently it's an implicit dependency
and so vunerable to races at build time.

(From OE-Core rev: 558705735aa9a2d640d1114bd809ca4ea7f0130d)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

insane.bbclass and friends: Fix sanity checks and multlib headers for n32

The n32 architecture is odd, in that it's a mips64 ABI which happens
to be 32-bit. To handle this, we need something in the environment
which can be used to distinguish it. The obvious place to stash this
is the ABI suffix, so we use "n32" as an ABI suffix. This allows
a couple of improved checks:

1. In insane.bbclass, we can use "linux-gnun32" to discern that it's
okay for a mips64 binary to be a 32-bit binary in some cases.
2. In multilib_header, we can check for the n32 ABI, and use a distinct
value.
3. In siteinfo, add linux-gnun32 as a synonym for linux, similar to
what's done for linux-gnux32, and tell the mips*-linux-gnun32 variants
to pick up the corresponding mips-linux site configs.

Note that the multilib header wrapper already has n32 hooks in it, there
was just nothing creating -n32 header variants.

(From OE-Core rev: c8e8e8ba22eaa335ac72f0e5b317f804035133e2)

Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk_base: Ensure that the multilib cross-canadian tools are used

Update the host toolchain list, for cross-canadian toolchains, to ensure
that all of the supported multilibs are built and installed.  This
dynamically generates the dependnecy set based on the current multilib
configuration.

(From OE-Core rev: 54bc658416ea5679bbfdc76e3ef8767c0a15211c)

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk_base: Update extraction script for multilibs

When multilibs are enabled, there will be more then one environment
file created.  We need to be sure to process each environment file.
The next function can simply use the last environment file processed
to get the magic value(s) that it requires.

(From OE-Core rev: 6f0537c835c35dcff5154de0bec066ec3e71a4f8)

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

multilib - crosssdk: Stop building multilib for crosssdk packages

Crosssdk packages are not actually multilib packages, so treat them
the same as other nativesdk packages in the multilib, base, and
classextend components.

(From OE-Core rev: 15834451525453e0f7ceac25d4f98117f1825f37)

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

multilib: Add support for cross-canadian multilib packages

Add support for the generation of cross-canadian packages.

Each cross-canadian package has:
 PN = "pkg-cross-canadian-${TRANSLATED_TARGET_ARCH}"

in order for that to be evaluated properly with multilibs enabled, it was
necessary to detect both the presence of the cross-canadian packages and
then update the vars using the OVERRIDE for the multilib.  Additional checks
were made to ensure that any dependency that sais "cross-canadian" did not
get prefixed with the MLPREFIX.

Also, make sure that even when building multilib cross-canadian packages,
we only use the single SDK PACKAGE_ARCH, we don't want or need variants.

(From OE-Core rev: 132a182e2f6c330aa645de42c1aeb386e43bddd3)

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

zeroconf: remove

We already have avahi in OE-Core and some cursory research suggests that
avahi is preferred over this package, which has apparently not seen a
release since 2006. Nothing in OE-Core actually refers to it, so let's just
remove it.

CC: Cristian Iorga <cristian.iorga@intel.com>
(From OE-Core rev: 88b4ec0b8c7c75b8570fc201c705937b459bb43e)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake: hob/builder: Hob crashes because of bad init

Image selection is not properly initialized to none,
and it used before having a chance to have a value.
Due to dynamic nature of Python, variable is used before
it exists, in this case. This causes a crash.
Bug introduced during the fix of [YOCTO #3228]

Fixes [YOCTO #3334]

(Bitbake rev: 1c540541c5397c38dca880a79df9ebfcda576d4c)

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev: upgrade to 182

This is the final upgrade of udev. Futher upgrades will only
come in conjunction with systemd.

The v4l1 removal patch is deprecated as the bug is fixed inside udev.
There is a new patch fixing the path for default sh interpreter.
New debug binaries are generated, and udev.inc is modified to package
those correctly.
The install locations changed for udevd and udevadm, so the scripts
are updated accordingly.

(From OE-Core rev: 3cbe52b94c4d559a037347ac419fafee5af84fe6)

(From OE-Core rev: 8fc73baecf1b21b1a3e7eff478e25d2a7cae2879)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Conflicts:

meta/recipes-core/udev/udev_164.bb

sgw - Fixed up DEPENDS += and added some OECONF options that where in the
meta-oe version and make sense to be included.

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

diffutils: fix rebuilds

If diffutils rebuilds it tries to remove with "rm" files that don't exist
anymore, resulting in an error.

Use rm -f so the removal always succeeds.

(From OE-Core rev: becd38412a95f3f9f6c3450a87a7204be032d2e6)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

perl: update dependencies

Update dependencies for perl modules again. When only install
perl-module-file-glob, run perl script with "require File::Glob;" will
fail. Update dependencies to fix that.

[Yocto 3069]

(From OE-Core rev: 1554e690d8d074f3bbe484d2acfebde4b94e3738)

Signed-off-by: Kang Kai <kai.kang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cogl/clutter: Explicitly depend on libdrm for GLX

cogl and clutter explicitly rely on libdrm being present when using the glx
backend. If its not listed in DEPENDS and an alternative to mesa is used, it
may not actually be present. This patch ensures it is and fixes a build
race condition which could see dependencies like clutter-box2d failing to
compile due to missing pkgconfig dependencies.

(From OE-Core rev: afb3ee76cef109c7ba4a760d834839ef277e30fc)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob/settings: add a scroll bar for the box with mirrors

- added a scroll bar in the shared state tab from settings
- added a signal on it, so when you add a new mirror, it will
auto-scroll to the end of the list

[YOCTO #3229]
(Bitbake rev: 00afd6a25c0cc0a4fcddd9f7c26a17ef6c47cbd2)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake: hob/builder: Image selection is remembered while changing adv settings

Image selection is remembered correctly even after
advanced settings have been changed.
Selected image was reset even when it was not the case.

Fixes [YOCTO #3228]

(Bitbake rev: 93fb0a2c56100b2bbc8769af9ae2343c05e72193)

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob/builddetailspage: set "Log" page as default

- when you do a build you should see "Log" page, not
"Build configuration" page

[YOCTO #2569]

(Bitbake rev: 431cb80d4d5222f832f6141b8578291f2f14a131)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob/imageconfigurationpage: progress bar shows when recipe parsing is stopped

-when the recipe parsing process is stopped, the progress bar shows
"Stopping recipe parsing"

[YOCTO #3259]
(Bitbake rev: d20626bd717bb8f5cfd73b91337af880198db247)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake: hob/hobwidget: "Brought in by" column is now displayed correctly

In "Edit Recipes" and "Edit packages" pages, the "Brought in by"
column is displayed correctly, with the right number of additional
packages and a proper title.

Fixes [YOCTO #2195].

(Bitbake rev: 4d1d3e5a54eb718e2eee02f734d929f15ccf99ce)

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: stop build without percentage shown

added a method in progressbar.py that doesn't update the
percentage in the progress bar shown in hob.

the call of the method is done in builder.py.

(Bitbake rev: 7ab5775fceda1055b86bdc3313fc4bf928bf5155)

Signed-off-by: Andrei Dinu <andrei.adrianx.dinu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: change behavior for "cancel" button from the Recipe selection page

-when returned to the Image configuration page, after canceling on the
Recipe selection page, the image selected previously is now shown corectly

[YOCTO #3205}
(Bitbake rev: 0a755026661a18ae386eb64b807e9e9e8f0dfe4c)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: image size displayed wrong size in edit packages screen

    size_str = '%.1f' % (size*1.0/(1024*1024)) + ' MB'
    from /bitbake/lib/bb/ui/crumbs/hobpages.py file transformed
    the size in MB. In our file it was again multiplied by 1024
    instead of doing a division by 1024, which brought a faulty size on
    the edit packages screen.

(Bitbake rev: 7dcea3884a45973ae332695dc8a53814b701151f)

Signed-off-by: Andrei Dinu <andrei.adrianx.dinu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: event/hob: Add a button for network tests in the proxy settings

This patch allows the user to check the network connectivity in
the "Proxy" page ("Settings" dialog) by adding a button which provides
this functionality. It also disables retrigerring sanity checks if the
proxy values are changed, since now the proxy checks are explicit.
Note that this patch depends on a patch in oe-core
("sanity.bbclass: trigger network tests explicitly"). It will
not work properly if the patch in oe-core is not merged.

[YOCTO #3026]

(Bitbake rev: cb1354d29c0be27aee57b9783c724457ef6725fb)

Signed-off-by: Bogdan Marinescu <bogdan.a.marinescu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

automake: update dependencies

Remove the RDEPENDS for nativesdk because the nativesdk-automake also
needs perl modules.

Add dependencies perl-module-thread-queue and perl-module-threads.

Remove redundant dependencies that they are already required by autoconf
and autoconf is required by automake.

In this removed list, "-->" present "required by":

  perl-module-cwd --> perl-module-file-path --> autoconf
  perl-module-dynaloader --> perl-module-xsloader --> perl-module-fcntl
     --> perl-module-file-stat --> autoconf
  perl-module-exporter-heavy --> perl-module-exporter --> autoconf
  perl-module-constant --> autoconf
  perl-module-errno --> autoconf
  perl-module-file-basename --> autoconf
  perl-module-file-compare --> autoconf
  perl-module-file-copy --> autoconf
  perl-module-file-glob --> autoconf
  perl-module-file-spec-unix --> perl-module-file-spec
         --> perl-module-io-file --> autoconf
  perl-module-file-stat --> autoconf
  perl-module-getopt-long --> autoconf
  perl-module-io --> perl-module-IO-handle --> perl-module-IO-seekable
     --> perl-module-io-file --> autoconf
  perl-module-io-file --> autoconf
  perl-module-posix --> autoconf

Bump up PR.

(From OE-Core rev: cd15622712c517dc72242c1066ca6eb4bc5094a8)

Signed-off-by: Kang Kai <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

autoconf: update runtime dependencies

Update autoconf runtime dependencies on perl and perl modules. And
remove RDEPENDS for nativesdk because the nativesdk-autoconf has same
dependencies with autoconf.

Then fixes autoreconf runs failed both on target and toolchain.

Bump up PR.

[Yocto 3100]

(From OE-Core rev: 19a4d6498b262a53456f43fabb66d821014d2656)

Signed-off-by: Kang Kai <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

perl: update dependencies among modules

Run autoreconf fails because it uses several perl modules and they
requires other perl modules. So update these dependencies for:
  perl-module-exporter
  perl-module-file-glob
  perl-module-file-path
  perl-module-file-spec
  perl-module-file-stat
  perl-module-io-file
  perl-module-io-handle
  perl-module-io-seekable
  perl-module-posix

And RDEPENDS rules in file perl-rdepends_5.14.2.inc don't work for
nativesdk perl module packages. Replace all "perl" with "${PN}" in the
file to fix that.

In nativesdk.bbclass it calls
oe.classextend.NativesdkClassExtender().map_packagevars() to map package
vars include var RDEPENDS. In map_packagevars():

  for pkg in (self.d.getVar("PACKAGES", True).split() + [""]):

the value of var "PACKAGES" may not be calculated correctly, so for
all the nativesdk packages created by

  PACKAGES_DYNAMIC_virtclass-nativesdk += "^nativesdk-perl-module-.*"

dependencies are wrong.

This is similar with 51cbb5ae76a22d465e2f6c5ef923ec2682624e3b.

Bump up PR.

(From OE-Core rev: c1f1368a680ae596e4d974a2cbbd253abc5118f8)

Signed-off-by: Kang Kai <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sato-icon-theme: merge sato-icon-theme.inc into recipe

nothing else needs sato-icon-theme.inc

(From OE-Core rev: b6bbfcfa329ea2695f1da53f880ae1f2cf2cf2e1)

Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>