acl: Revert most of my last change, just fix RPATH QA

Signed-off-by: Tom Rini <tom_rini@mentor.com>

schroedinger: Fix RPATH QA errors

Signed-off-by: Tom Rini <tom_rini@mentor.com>

libcdio: Fix RPATH QA errors

Signed-off-by: Tom Rini <tom_rini@mentor.com>

kernel.bbclass: fix retrieving dependencies between modules.

get_dependencies() searched for a filename key in dependencies
directory with the /lib/modules/kernel_version prefix while
the directory had entries stripped of it.

Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Acked-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

Add flex to the linux-atm dependencies

As the linux-atm tools also link against the flex libraries (-lfl) the
runtime is also needed as a dependency.

Signed-off-by: Alex Bennee <ajb@cbnl.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

udev_151/162/165.bb: add missing PACKAGE_ARCH for nokia900

Signed-off-by: Klaus Kurzmann <mok@fluxnetz.de>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

utils.py: use bb.utils.vercmp_string instead of bb.vercmp

* Fixes deprecation warning with BitBake master.

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
CC: Chris Larson <chris_larson@mentor.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

tiff: Fix RPATH QA issues

Based on the fix for 'acl'

Signed-off-by: Tom Rini <tom_rini@mentor.com>

acl: Fix RPATH QA issue and clean up

We take some bits from poky to fix the RPATH QA issue and while in here
clean it up a bit based on Ubuntu/Debian.

Signed-off-by: Tom Rini <tom_rini@mentor.com>

cacao-native: modell the package version after the git recipes

* bump PR

bug-image-production: Add back bash-sh package

bash: Add bash-sh package to switch /bin/sh to /bin/bash

An old change from hrw in the BugLabs repo I forgot to bring over earlier.

org.freedesktop.dbus: modify recipe so localization/compilation occurs correctly.

kaeilos.inc: switched to eglibc

libpng.inc: Package library in libpng and provide libpng3

Signed-off-by: Khem Raj <raj.khem@gmail.com>

contrib/testing/testscript.sh: Dont define BBMASK

* BBMASK is set in bitbake.conf by default and it
  skips nonworking and obsolete recipes which is
  what we want emptying it here could cause redundant
  parse of the graveyards

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Tom Rini <tom_rini@mentor.com>
Signed-off-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>

glib-2.0 2.28.0: Bring back glib-mkenums-interpreter.patch

Signed-off-by: Tom Rini <tom_rini@mentor.com>

bitbake.conf: export LC_ALL = "C"

* thus the build system language won't create troubles during builds
(example : perl is hacking some generated files' content but only
supports english outputs).

* this problem is only met when using bitbake master

Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

gnutls.inc: Add --disable-rpath

Signed-off-by: Tom Rini <tom_rini@mentor.com>

openssl: Update from 1.0.0c to 1.0.0d

Security fixes only so just bumping everyone that pinned up.

Signed-off-by: Tom Rini <tom_rini@mentor.com>

sqlite3-native: Fixup DEPENDS

The way things were before we were missing tcl-native.  In addition,
we can clean up the recipes slightly since sqlite3.inc is where we
set BBCLASSEXTEND = "native"

Signed-off-by: Tom Rini <tom_rini@mentor.com>

gettext: Pass --disable-rpath on target

We disable rpath stuff as this casues problems with libtool 2.2
and 2.4 is fine.

Acked-by: Marco Cavallini <m.cavallini@koansoftware.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

db: Bring in 5.1.19 from poky

Based on e77207ccdf15f34651a086099ed4816712e3656b in poky, bring in
5.1.19 (no RPATH issues when using libtool 2.2) and drop un-pinned
4.3.29 and 5.0.21 as this is backwards compatible.

Acked-by: Marco Cavallini <m.cavallini@koansoftware.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

com.buglabs.bug.base: add support for bug 1.x

Without that fix, it fails to fetch the now gone hirop branch

Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@no-log.org>
Acked-by: Stefan Schmidt <stefan@datenfreihafen.org>

task-bug: make it compatible with bug 1.x,machine arch ${PN}

Some packages like bug2v4l2 are specific to bug 2.0, so we need
  to machine arch the package depending on them.

Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@no-log.org>
Acked-by: Stefan Schmidt <stefan@datenfreihafen.org>

bug20: add MACHINE_EXTRA_RRECOMMENDS

Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@no-log.org>
Acked-by: Stefan Schmidt <stefan@datenfreihafen.org>

hostapd 0.7.3: convert to libnl2

Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@no-log.org>
Acked-by: Stefan Schmidt <stefan@datenfreihafen.org>

pingus 0.7.2: fix newer boost(1.45.0) detection (break older boosts)

Without that fix we have:
  Fatal Errors:
    * library 'boost_signals' not found

  FATAL: scons build execution failed.

Note that this commit is incompatible with older(than 1.45.0) boost.

From #openmoko-cdevel on Freenode on irc:
  Feb 12 11:43:07 <JaMa|Off>      GNUtoo|laptop: that boost version is default,
  so if you fix it for default version and break for older it's not so bad (
  distros using older boost should handle it)
  [...]
  Feb 12 11:48:37 <GNUtoo|laptop> JaMa|Off, can I fix like that or should I send
  a patch for breaking older boosts and distros?
  Feb 12 11:49:35 <JaMa|Off>      GNUtoo|laptop: nobody in OE is pinning boost..
  so just note it in commit (that it won't work with old boost anymore)

The patch is kept in tree because it could be usefull if:
 * Someone is still using old boost
 * -mt versions are re-added in the future

Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@no-log.org>

linux-bug: fix undefined reference to `__gnu_mcount_nc' issue and bump PR

Without that fix we have:
    LD      .tmp_vmlinux1
  init/built-in.o: In function `run_init_process':
  calibrate.c:(.text+0x10): undefined reference to `__gnu_mcount_nc'
  init/built-in.o: In function `init_post':
  calibrate.c:(.text+0x3c): undefined reference to `__gnu_mcount_nc'
  init/built-in.o: In function `do_one_initcall':
  calibrate.c:(.text+0x148): undefined reference to `__gnu_mcount_nc'
  init/built-in.o: In function `name_to_dev_t':
  calibrate.c:(.text+0x2cc): undefined reference to `__gnu_mcount_nc'
  init/built-in.o: In function `set_reset_devices':
  calibrate.c:(.init.text+0x10): undefined reference to `__gnu_mcount_nc'
  init/built-in.o:calibrate.c:(.init.text+0x38): more undefined references to `__gnu_mcount_nc' follow
  make: *** [.tmp_vmlinux1] Error 1
  FATAL: oe_runmake failed

Thanks khem in #oe on Freenode on irc for help:
  Feb 14 19:21:26 <khem>  GNUtoo|laptop: are you using profiling
  [...]
  Feb 14 19:57:13 <khem>  GNUtoo|laptop: ok so here it your problem you are passing -pg
  Feb 14 19:57:16 <khem>  to compiler
  [...]
  Feb 14 19:58:11 <khem>  GNUtoo|laptop: thats only useful if you will provide mcount and other profiling hook implementation
  Feb 14 19:58:16 <khem>  compiler sort of expects it

  The solution I chose was to disable the CONFIG_FTRACE that added -pg (profiling compiler option)

Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@no-log.org>

busybox: remove 1.18.2

now that 1.18.3 is the preferred for Ångström and SlugOS

Signed-off-by: Eric Bénard <eric@eukrea.com>
Acked-by: Khem Raj <raj.khem@gmail.com>

slugos: prefer busybox 1.18.3

Signed-off-by: Eric Bénard <eric@eukrea.com>
Cc: Mike Westerhof <mike@mwester.net>
Acked-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Mike Westerhof <mike@mwester.net>

angstrom-2010: prefer busybox-1.18.3

this is the latest stable version

Signed-off-by: Eric Bénard <eric@eukrea.com>
Cc: Koen Kooi <koen@openembedded.org>
Acked-by: Koen Kooi <koen@openembedded.org>

busybox: add 1.18.3 and its fixes

- 1.18.3 is 1.18.2 + the fixes already presents in OE
- additional fixes from http://busybox.net/downloads/fixes-1.18.3/
were added to the recipe

Signed-off-by: Eric Bénard <eric@eukrea.com>
Acked-by: Khem Raj <raj.khem@gmail.com>

busybox: add group to -passwd package

Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
Signed-off-by: Eric Bénard <eric@eukrea.com>
Acked-by: Tom Rini <tom_rini@mentor.com>
Acked-by: Khem Raj <raj.khem@gmail.com>

nodejs: add 0.4.0

compile tested for armv5, previous node-cross-cc_0.2.6.patch
doesn't seems anymore necessary unless I made a mistake, AJ may
you please check this point ?

Signed-off-by: Eric Bénard <eric@eukrea.com>
Cc: AJ ONeal <coolaj86@gmail.com>
Acked-by: Khem Raj <raj.khem@gmail.com>
Tested-by: AJ ONeal <coolaj86@gmail.com>

gnokii: remove old recipes

Signed-off-by: Eric Bénard <eric@eukrea.com>
Acked-by: Khem Raj <raj.khem@gmail.com>

gnokii: add latest version

- build tested for armv5 with PARALLEL_MAKE="-j8" so I removed
PARALLEL_MAKE = "" which was set in previous recipes.

Signed-off-by: Eric Bénard <eric@eukrea.com>

gnokii-0.6.14: fix download link

Signed-off-by: Eric Bénard <eric@eukrea.com>
Acked-by: Khem Raj <raj.khem@gmail.com>

util-linux-ng: add 2.18

from the changelong, the most important update over 2.17
is a memory leak fix in mount

Signed-off-by: Eric Bénard <eric@eukrea.com>
Acked-by: Khem Raj <raj.khem@gmail.com>

angstrom 2008: prefer a libsoup from the stable series

Signed-off-by: Koen Kooi <koen@openembedded.org>

angstrom next: move to glib 2.28.0

Signed-off-by: Koen Kooi <koen@openembedded.org>

libsoup-2.4 2.33.6: lower default pref

* it's a release from the unstable series
* it needs a too new glib-2.0

Signed-off-by: Koen Kooi <koen@openembedded.org>

gcc-4.5: Bring latest from linaro 4.5 and bump svn SRCREV for upstream

gcc-4.5-linaro-r99474.patch is disabled because it ends up
bombing x86 cross-gcc build see
http://lists.linaro.org/pipermail/linaro-dev/2011-February/002614.html
gcc-4.5-linaro-r99474.patch is disabled because it causes GCC crash
for XScale see
http://lists.linaro.org/pipermail/linaro-dev/2011-February/002600.html

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Martin Jansa <Martin.Jansa@gmail.com>

readline 6.1: Move 'p2' into PV and out of PKGV

Otherwise we have problems with readline.

Signed-off-by: Tom Rini <tom_rini@mentor.com>

ncurses: Move ${PATCHDATE} into PV rather than just PKGV

This was making explode_deps unhappy.

Signed-off-by: Tom Rini <tom_rini@mentor.com>

autotools: symlink where we can

Signed-off-by: Chris Larson <chris_larson@mentor.com>
Acked-by: Khem Raj <raj.khem@gmail.com>

autotools: split out a oe_autoreconf function

This functionality logically belongs in its own function, and it makes it
easier to explicitly run against subdirs in certain cases.

Signed-off-by: Chris Larson <chris_larson@mentor.com>
Acked-by: Khem Raj <raj.khem@gmail.com>

autotools: cleanup

Signed-off-by: Chris Larson <chris_larson@mentor.com>
Acked-by: Khem Raj <raj.khem@gmail.com>

native: fix duplicating virtclass-native entries in OVERRIDES

A function was called from ${@} in OVERRIDES which was supposed to return the
bits to prepend when in virtclass context, yet it *set* overrides instead,
returning None.  This resulted in 1) adding an extra virtclass-native to
OVERRIDES each time it was expanded, and 2) appending None, causing
'localNone' to be in overrides rather than the expected 'local'.

Signed-off-by: Chris Larson <chris_larson@mentor.com>

bt-gps: fix hardcoded paths, fix QA issues, use bindir/datadir variables

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

task-shr-feed: add bison and automake

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

libsoup-2.4: add 2.33.6 version needed for newer webkit-efl

Acked-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Klaus Kurzmann <mok@fluxnetz.de>
Tested-by: Klaus Kurzmann <mok@fluxnetz.de>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

SHR: drop P_V for glib-2.0 (upgrades from 2.26.1 to 2.28.0)

Acked-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Klaus Kurzmann <mok@fluxnetz.de>
Tested-by: Klaus Kurzmann <mok@fluxnetz.de>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

glib-2.0: add 2.28.0, needed by newer libsoup

Acked-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Klaus Kurzmann <mok@fluxnetz.de>
Tested-by: Klaus Kurzmann <mok@fluxnetz.de>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

openmoko-gps: bump SRCREV, fix QA issues, fix hardcoded path, use bindir/datadir and rename recipe

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

mokoeightball: fix hardcoded paths and QA issues

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

cdparanoia (svn): use build variables to fix packaging when using empty prefix

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>

linux-2.6.24: phy handling fixed for hipox machine

* phy switching enabled
* link status handling fixed

Signed-off-by: Mario Schuknecht <m.schuknecht@dresearch.de>
Acked-by: Steffen Sledz <sledz@dresearch.de>

task-base: bump PR to pick changes from nokia900.conf

* lack of coffee made me thing that MACHINE_KERNEL_PR bump in
  648bd1c785fae9de66001dc588720366b9ebda2c is what I need, sorry for
  that

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

qt4e-demo-image: Switch to using task-qt4e.bb

In task-qt4e.bb put the xmlpatterns bits in RRECOMMENDS for QT 4.6.x

Signed-off-by: Tom Rini <tom_rini@mentor.com>

task-qt4-x11: Move libqtxmlpatterns4 to RRECOMMENDS

Fixes qt4 4.6.x images.

Signed-off-by: Tom Rini <tom_rini@mentor.com>

Revert "Remove some patches, which was moved by previous linux patch"

I pulled this patch in prematurely which broke the at91 machines,
so back this out.  Sorry folks!

This reverts commit cfa8a2a8b291cb0a3426a3120ce31e6e74970e37.

Revert "gcc-4.5: Bring latest from linaro 4.5 and bump svn SRCREV for upstream"

This reverts commit f76026a6c1a00bd0b2f1ae2b045dee995293edc6.

Pushed prematurely

Signed-off-by: Khem Raj <raj.khem@gmail.com>

ti-pru-sw: Update SRC_REV to 24

For the ti-pru-sw-examples, ti-pru-sw-app-loader, ti-pru-sw-edma-driver,
ti-pru-sw-edma-library:

* Update SRC_REV to 24

Signed-off-by: Melissa Watkins <m-watkins@ti.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

bitbake.conf: Fix wrong comparison

The expression ('' or 'custom') will always result in 'custom' so the expression will become False when the variable (TARGET_OS, SDK_OS) is '' while its expected to be True in this case also.

Signed-off-by: Imran Mehmood <imran_mehmood@mentor.com>
Acked-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

nokia900: add kernel-module-ubifs to RRECOMMENDed modules

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

testbuilders: change oestats server to tinderbox.openembedded.org

vala: 0.11.4.2 -> 0.11.6.1

binutils_2.21: Fix QA errors due to wrong RPATH

Signed-off-by: Khem Raj <raj.khem@gmail.com>

gcc-4.5: Bring latest from linaro 4.5 and bump svn SRCREV for upstream

Please test this patch out in your respective combinations and report
and regressions you see.

Signed-off-by: Khem Raj <raj.khem@gmail.com>

u-boot 2009.01: Fix 'weak' issues

Based on existing patches for other boards.

Signed-off-by: Tom Rini <tom_rini@mentor.com>

libgisi: packaging++

flac: fix build problems with rpath on older libtool

* bump INC_PR

spojegui: bump SRCREV and remove upstream applied desktop file fixes for QA issues

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

alsa-state: enable HDMI audio by default

Signed-off-by: Koen Kooi <koen@openembedded.org>

libsdl-mixer: fix compiling for angstroem 2008.1

* libtool don't honors AC_CONFIG_AUX_DIRS in this
  case so not all libtool stuff is deleted, fixed it
  by copy the stuff from the build-script dir
  and remove the dir afterwards
* bump PR

autofs: set license field to GPLv2+

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>

bug-desktop-icons: Fix QA error in desktop files.

Its not allowed to use the file extension when not using a full path. Just use
the image name here.

rootfs_ipk: don't emit error code when ONLINE_PACKAGE_MANAGEMENT is really none

* currently do_rootfs fails when it's "none", because that test returns
  exit code '1':
  | log_check: Using /OE/tmpdir-shr/work/spitz-oe-linux-gnueabi/initramfs-kexecboot-klibc-image-1.0-r0/temp/log.do_rootfs.21324 as logfile
  | Logfile is clean
  | + rm -rf /OE/tmpdir-shr/rootfs/initramfs-kexecboot-klibc-image-tmp
  | + '[' none '!=' none ']'
  NOTE: package initramfs-kexecboot-klibc-image-1.0-r0: task do_rootfs: Failed
  ERROR: Function 'do_rootfs' failed (see /OE/tmpdir-shr/work/spitz-oe-linux-gnueabi/initramfs-kexecboot-klibc-image-1.0-r0/temp/log.do_rootfs.21324 for further information)
  ERROR: Task 1356 (/OE/dev/recipes/images/initramfs-kexecboot-klibc-image.bb, do_rootfs) failed with exit code '1'
  ERROR: '/OE/dev/recipes/images/initramfs-kexecboot-klibc-image.bb' failed

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

linux.inc: Add lzma-native for nokia900

Add a check for nokia900 to the existing DEPENDS mangler to add
lzma-native, otherwise:
|   LZMA    arch/arm/boot/compressed/piggy.lzma
| /bin/sh: lzma: command not found

Signed-off-by: Tom Rini <tom_rini@mentor.com>

Add a small helper script listing all recipes with QA errors.

Signed-off-by: Andreas Mueller <schnitzeltony@gmx.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>

Revert "gst-rtsp.inc: Add check to DEPENDS"

check is already in the DEPENDS list plus it had typo in it

This reverts commit 393c0d5bc5850592b51013d0918d0fe623d40de8.

python_2.6.6.bb: Fix QA ERRORS due to staging libdir in rpath

Signed-off-by: Khem Raj <raj.khem@gmail.com>

gst-rtsp.inc: Add check to DEPENDS

Signed-off-by: Khem Raj <raj.khem@gmail.com>

check_0.9.5.bb: Add patch to use quadrigraphs in macro names

Signed-off-by: Khem Raj <raj.khem@gmail.com>

package_deb.bbclass: remove double DISTRO_PR from the package version.

PKGR already contains DISTRO_PR, so no need to add it twice.

Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

package_deb.bbclass: depend on apt, not dpkg.

Both package_deb.bbclass and rootfs_deb.bbclass use apt-get.

Without apt system cannot be bootstrapped because there is no optimal
way to fetch new packages. Apt depends on dpkg, so switching from dpkg
to apt in EXTRA_DEPENDS provides both.

Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

u-boot: Make sure we use override HOSTCC for env tools

Upstream says that the right way to build a target binary of the env
tools is to pass in HOSTCC=your-cross-cc.  Bump the PR in u-boot_git.bb
to fix as many easily fixable ones as possible, but other platforms
may have been shipping bad binaries (since we had skipped testing
on this package due to GNU_HASH issues when built for the target).

Signed-off-by: Tom Rini <tom_rini@mentor.com>

Revert 6a7f8c in favor of a real fix

Signed-off-by: Chris Larson <chris_larson@mentor.com>

package_deb.bbclass: Use PKGV/PKGR variables where applicable.

Using PV/PR causes problems like libncursesw5_5.7-r16 depending on
libtinfo5 (>= 5.7+20110115).

Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

python: generate-manifest-2.6.py, update few entries

* config was in python-dev as well as python-distutils, keep in
  distutils
* linecache.* was in python-core as well as python-lang, keep in core

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

python: generate-manifest-2.6.py, more s/python/${PN}/g and regenerate it

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

python: generate-manifest-2.6.py, sort packages when generating output file

* iteritems and packages didn't have defined order, so every regeneration created huge patch, hard to review
* alphabetical order should work fine here as long as we have disjunct FILES

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

python: merge few s/python/${PN}/g from poky

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

python: upgrade from 2.6.5 to 2.6.6

* refresh patches from poky where applicable

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

packaged-staging: Don't pass a trailing ' ' to bb.mkdirhier

Signed-off-by: Tom Rini <tom_rini@mentor.com>

image.bbclass: move insert_feed_uris() to rootfs_ipk.bbclass

This function generates feed configuration for opkg and as such
should be used only for images using ipk, not rpm or deb.

Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>

utils: drop exported var env caching for oe_popen

This was dangerous, as an oe_run executed from the configuration metadata
would result in caching the config metadata information, rather than the final
metadata information.  We could also cache it by the datasmart instance, but I
don't think its worth the trouble at this time, particularly given the next
bitbake release exports all exported vars into the process environment
automatically, avoiding the need for any of this.

Signed-off-by: Chris Larson <chris_larson@mentor.com>

packaged-staging: add debug info for oe_run failures

Signed-off-by: Chris Larson <chris_larson@mentor.com>

deb packages support: switch from /var/dpkg to /var/lib/dpkg

OE generally tries to use /var/dpkg, but the change is missing
in some places, so it is needed to move completely to one of the
options. Keeping the dpkg's default /var/lib/dpkg causes less
maintenance overhead for OE.

Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>