bitbake: fetch2: Fix missing function call

Fix the issue:

File: '/srv/home/pokybuild/yocto-autobuilder-new/yocto-slave/build-appliance/build/bitbake/lib/bb/fetch2/__init__.py', lineno: 813, function: try_mirror_url
     0809:    except bb.fetch2.BBFetchException as e:
     0810:        if isinstance(e, ChecksumError):
     0811:            logger.warn("Mirror checksum failure for url %s (original url: %s)\nCleaning and trying again." % (newuri, origud.url))
     0812:            logger.warn(str(e))
 *** 0813:            self.rename_bad_checksum(ud, e.checksum)
     0814:        elif isinstance(e, NoChecksumError):
     0815:            raise
     0816:        else:
     0817:            logger.debug(1, "Mirror fetch failure for url %s (original url: %s)" % (newuri, origud.url))
Exception: NameError: global name 'self' is not defined

(Bitbake rev: 5f960017bb1d4274a5bd5b598971f38c12fc0f11)

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

bitbake: server/bitbake: Remove launchUI method

With the removal of the none server type, we can remove the launchUI
method and simplify the code slightly.

(Bitbake rev: 9bef2f2dd0bcaa59528ebcb3c1ce053b7dff1ec6)

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

bitbake: server: Remove none server

The process server backend has been serving well as the default for a long
time now and the UI model is much better thought out that it used to be. With
the move to make bitbake a memory resident process, the none server is now
looking rather pointless and complicates the code needlessly. Lets therefore
now remove it.

(Bitbake rev: 9af03a89605e3db9bce3cea1e0f2d0b6cfaa6fe1)

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

yocto-bsp: re-enable AutoAddDevices in xorg.conf for generated BSPs

This was added quite a long time ago because of poor interactions
between HAL and the X server when it came to enabling input devices.
HAL is long gone and I think it's safe to say we don't need to disable
this any longer, especially as it gets in the way of being able to plug
in the keyboard/mouse after boot.

(From meta-yocto rev: e06ab1e030e8cfbc259500b1a0b958fe752fb872)

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

bind: add ipv6 support

Make ipv6 configure dependent on DISTRO_FEATURES.

(From OE-Core rev: 984eb9e25b658c5e9a870983841060aca32bc137)

Signed-off-by: Li Wang <li.wang@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

util-linux: Add ability to compile with nativesdk

Some of the tools in the util-linux are used for disk and text file
operations in the nativesdk so as to get around different versions
that may exist on the host system.

(From OE-Core rev: 2348ce4fccf0ec4f3bc7aacf953eb03dfac0642a)

Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libpam: Avoid wildcards in the SRC_URI

Remove the wildcard from the SRC_URI.  This causes problems when you .bbappend
and add a FILESEXTRAPATHS entry.  The unpack task may be unable to find the
files to unpack leading to an error.

Avoid wildcards at all costs...

(From OE-Core rev: 6d3705123dd2f808a9778326aa04a2854f7b5378)

Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cmake.bbclass: modify construction of compiler flags

Use CFLAGS instead of CPPFLAGS for C_FLAGS variants.

When debug optimization is enabled in the local.conf, the debug (-O0) vs
production (-O2) does not change in the builds.  As the CPPFLAGS do not
contain the optimization settings.

Also the CXX_FLAGS are based on CXXFLAGS, so it makes sense to similarly
set the C_FLAGS based on CFLAGS.

(From OE-Core rev: 558662927be550aeb8dd163f65e16b1750bbd127)

Signed-off-by: Joe Slater <jslater@windriver.com>
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dpkg-native: Fix native perl path

The path to the native perl was incorrect leading to rootfs failures. This
patch corrects that problem, it's a complementary fix for commit:04432446.

(From OE-Core rev: 0f99d7fed094a59d2c5c01c83ea38dc852aadf6b)

Signed-off-by: Ming Liu <ming.liu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

guile: don't search for libreadline in host libdir

Fix the QA warning:
WARNING: guile: The compile log indicates that host include and/or library paths were used.

(From OE-Core rev: 1582975cd89b0b71c93913f07648c67f2b18bc99)

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

qmake_base.bbclass: Add linux-gnun32-oe-g++ to QMAKESPEC

For some mips targets, TARGET_OS is set to linux-gnun32, while
linux-gnun32-oe-g++ is not listed in the default QMAKESPEC list of
qmake in oe-core/wrlinux. This would cause build failures for qt
apps, so add a matching rule to fix this.

(From OE-Core rev: 70b75d506e6c4b46694b00d674df9d4a94140bd6)

Signed-off-by: Xin Ouyang <Xin.Ouyang@windriver.com>
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

siteinfo.bbclass: Add mips64 common siteinfo

There is no mips64-common file, replace the definition with mips-common.

(From OE-Core rev: b6107abd50da651596c43119001cfc80fdf87554)

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

consolekit: remove /var/run from package

The /var/run/ConsoleKit directory doesn't need to be included in the
package as it is created by console-kit-daemon if it doesn't exist.
The /var/run directory is already created by base-files.

(From OE-Core rev: a433b86d15321d5061f7bdb9a0f1b4d58de2129c)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

resolvconf: remove /var/volatile/run/resolvconf/interface from package

- Remove /etc/resolvconf/run/interface from package as it actually uses
/etc/resolvconf/run/interface
- Create /var/run/resolvconf/interface on startup using
populate-volatiles.sh and tmpfiles.d for systemd
- Create symbolic link from /etc/resolvconf/run to /var/run/resolvconf

(From OE-Core rev: 619d78016be8d47691e3b2d218a6b855364541a3)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cups: remove /var/run from package as cupsd will populate it on startup

The /var/run/cups and /var/run/cups/certs directories don't need to be
included in the package as they are created by cupsd with the proper
permissions if they don't exist. The /var/run directory is already
created by base-files.

(From OE-Core rev: af99c290a0f589a5cb1d6426c78804f2d99ae02a)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bind: remove empty /var/run directory

The /var/run directory is already created by base-files.

(From OE-Core rev: 533f34c4c4596efbd7798f0819bab9642cc97be2)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev: do not create symbolic link from /run to /var/run

The /run directory is now a tmpfs with /var/run as a symbolic link
pointing to /run.

(From OE-Core rev: 691593177aa78a56ce138f1041872bebca2aa056)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev: do not create /var/volatile/run in init script

Previously in udev init, /var/volatile/run had to be created after
mounting all tmpfs filesystems so that udevd can write to /var/run
(a symbolic link to /var/volatile/run). This is because udev is
started before populate-volatile.sh.

Now that /var/run is a symbolic link to /run (a tmpfs filesystem),
/var/volatile/run doesn't need to be created anymore.

(From OE-Core rev: d557f6615701c9f2f461a10c30de1d9572424266)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: remove old /var/run workarounds

Do not clear /var/run on startup as it is a tmpfs.
Do not create empty /var/run/utmp on startup as it no longer seems
needed for systemd-update-utmp-runlevel.service to start properly.

(From OE-Core rev: 0c357b53b2c0123feeedfc202491b39eb639bfa7)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: do not create symbolic link from /run to /var/run

The /run directory already exists as a tmpfs with /var/run as a symbolic
link pointing to /run.

(From OE-Core rev: ee8e5b3ddaae1d3ae473a3cea2ff60fcee5877a2)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: do not create /var/volatile/run and /var/volatile/lock

The directories are not needed anymore.

Previously:
/var/run  -> /var/volatile/run
/var/lock -> /var/volatile/lock

Now:
/var/run  -> /run
/var/lock -> /run/lock

(From OE-Core rev: b314519f31699946140c93da961ff79e5ee28ccd)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base-files: add /run directory from FHS 3 draft specification

This adds the /run directory from the Filesystem Hierarchy Standard 3.0
Draft [1] and refactors the filesystem as follows:
- Remove creation of /var/volatile/run
- Remove creation of /var/volatile/lock
- Remove symbolic link from /var/run to /var/volatile/run
- Remove symbolic link from /var/lock to /var/volatile/lock
- Add symbolic link from /var/run to /run
- Add symbolic link from /var/lock -> /run/lock
- Add /run to /etc/fstab for sysvinit compatibility

[1] http://www.linuxbase.org/betaspecs/fhs/fhs.html#runRuntimeVariableData

(From OE-Core rev: 0e326280a15b0f2c4ef2ef4ec441f63f55b75873)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

findutils: backport fixes for documentation build errors

(From OE-Core rev: e34257bab558384936ba73f78902a5185ed51c49)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cracklib: Allow byte order patch to work on older Linux hosts

Older hosts don't have the htobe* and be*toh functions defined.
Instead we fall back to checking the endian and calling bswap_*
directly.  This works on both old and new hosts.

(From OE-Core rev: 52c83ea977b0f95917ec81dff394454e1a9bd541)

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>

gdb-cross-canadian: use correct exec-prefix path for python

Incorrect exec-prefix path was given to gdb which leads to gdb
startup failure when SDK is not installed to its original destination.
Gdb relocates the exec-prefix path, so it will work for SDKs that
are installed to different location. PYTHONHOME env in no longer
neeeded for gdb.

[YOCTO #3839]

(From OE-Core rev: e77603324332b932c73c9e22ab65a0b9b7c17798)

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

binutils: add two AArch64 related backports to 2.23.2

Update required to pass "movi" related build errors when gcc-4.8 is
used.

libgcrypt, slang, mysql5 were failing like this:

| {standard input}: Assembler messages:
| {standard input}:316: Error: immediate value out of range 0 to 255 at operand 2 -- `movi v3.8b,-106'
| {standard input}:348: Error: immediate value out of range 0 to 255 at operand 2 -- `movi v3.8b,-8'
| {standard input}:352: Error: immediate value out of range 0 to 255 at operand 2 -- `movi v3.8b,-27'

(From OE-Core rev: 2489151dbfc8bc002d89ab199d457ab3794c54a8)

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

libdrm: fix packaging after last upgrade

* 2.4.44 adds 2 more tests kmstest and vbltest which got packaged
  to PN, causing PKG rename and breaking upgrade path
  * check_data_file_clashes: Package libdrm wants to install file /usr/lib/libdrm.so.2
         But that file is already provided by package  * libdrm2
  * check_data_file_clashes: Package libdrm wants to install file /usr/lib/libdrm.so.2.4.0
         But that file is already provided by package  * libdrm2
  put them to PN-tests where they belong

(From OE-Core rev: 2c39ca5a6744de58013e9e43b9f6cc4efa66ece9)

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

freetype: update to upstream version 2.4.12

LICENSE.TXT md5sum changed as there's mention about part
of the code being in public domain added.

no-hardcode.patch removed as upstream has no longer the
problematic code to patch at all.

(From OE-Core rev: 2bd4982a466d913767318a961b0c70bb453f7018)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

curl: update to upstream version 7.30.0

0001-Fix-NULL-pointer-reference-when-closing-an-unused-mu.patch now
part of upstream.

(From OE-Core rev: 2d79a2f88b6676847ef868d3cc6475bd643b28a3)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gtk+: update to upstream version 2.24.18

(From OE-Core rev: 4d5ac121db74ff662293d38ba453416b0a54d5fd)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

harfbuzz: update to upstream version 0.9.17

avoid_double_version_h.patch removed as the problems seems
to be fixed different way in upstream, and this patch was
now removing necessary header install.

(From OE-Core rev: 7827e27ec4cd67d3821839209a29e4649e864b93)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

autogen-native: upgraded to 5.17.4

(From OE-Core rev: 849d98a45b3f21f55bfecdd9daf494a632d53a48)

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

python-dbus: upgraded to 1.2.0

The license itself didn't modify, just some comments regarding the license.

(From OE-Core rev: ccf27bfe1269df74fa00b77447fecfe2368c0191)

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

git: upgraded to 1.8.3

(From OE-Core rev: 7d1ab4fe3bd78799b2e6401c6413d43655de348e)

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

settings-daemon: Update to git HEAD 9a99528b02255450db81176abd9bbcc1dab9a4c1

(From OE-Core rev: 7d45d5c666880e3242e1fe0dcd2b16d6c0ebdf4e)

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

libtasn1: Updated from 2.14 to 3.3

(From OE-Core rev: 689ac095b868b3ee975169257e114981f2593b4b)

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

apt: Updated from 0.9.7.8 to 0.9.8.1

(From OE-Core rev: 76b32bbe16c790ed81fffe37e67d80d1ec754e4e)

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

systemd: fix ordering for machineid and run-postinsts services

The remount-rootfs.service unit has been renamed to
systemd-remount-fs.service in systemd v183 and later.

The run-postinsts script writes to /var/log (a symbolic link to
/var/volatile/log), so systemd-tmpfiles-setup.service is added to After=
in run-postinsts.service to ensure /var/volatile/log is created before
running the script.

[YOCTO #4490]

(From OE-Core rev: 8b59ec4eb761d88445da94bb90aa2c5db0bbf365)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dropbear: a fix for hang in dropbearkey, built for x32

This commit fixes runtime hang of 'dropbearkey' utility, built for a x32
target abi system. The hang was observed while generating ssh keys, with
this command:
  dropbearkey -t dss -f private

The issue is fixed by changing the code, where 'long' in x86_64 mode is
assumed as 64bit quantity. With the x32 abi, the processor is in x86_64
mode, but the 'long' is a 32bit quantity. Hence the fix uses 'long long'
instead of 'long' to define/access 64bit data variables.

Fixes bug:
[YOCTO #4496]

(From OE-Core rev: 8f5bc47729edb8cb051d81e9ff1680cb8d2eca25)

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-firmware: Split 'ar3k' in a linux-firmware-ar3k package

(From OE-Core rev: ea6a8c9b0dee89b0b25d341ac57666cf090d6ed3)

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

systemd: forward port uclibc patches to 204

(From OE-Core rev: e33b5ddeb14e60f42ea709f41965e4763d4bf027)

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>

rm_work.bbclass: ignore failure for removing pseudo folder

When building over NFS2/3 server, removal of pseudo folders will fail in
some cases for there are files in it still used by pseudo daemon, thus
cause ".nfsXXXXX" files generated which can't be removed by clients. This
will lead rm_work task fo fail.

These failures could be safely ignored because ".nfsXXXXX" files would be
automatically cleared by NFS server when no clients keep opening them.

[YOCTO #4531]
[ CQID: WIND00412051 ]

(From OE-Core rev: 9742e866f545bc0d04aca697b541ed88f4e1764a)

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

systemd: update to 204

Removed patches integrated upstream.
Added backport of fix for utmp not running at shutdown.
Updated serial-getty service.
Added missing util-linux dependency.

(From OE-Core rev: 9e39d10ec92db23c9765b83559c32c0302bc8c5d)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate-volatile.sh: remove repeated leading slashes in TNAME

This avoids triple slashes in the generated /etc/volatile.cache to
reduce disk usage and in the output when verbose mode is enabled.

As all the paths for volatiles start with a slash, we can change
TNAME=${ROOT_DIR}/${TNAME} to TNAME=${ROOT_DIR}${TNAME}. To avoid
a double slash when ROOT_DIR is /, we strip the extra slash from
ROOT_DIR.

(From OE-Core rev: af56670f656ec0989aa7fd6cf6037cbc9cd88185)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dbus: remove /var/run from package as it is created on startup

/var/run is a tmpfs that is created on startup.
For sysvinit, /var/run/dbus is created by populate-volatiles.sh.
For systemd, /var/run/dbus is created implicitly by dbus.socket when
creating a listen stream socket at /var/run/dbus/system_bus_socket.

(From OE-Core rev: c11ba731fb245683148d0a8485b8c4d73bf94c28)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

binutils: the cat should always succeed when building config.cache

There is a race condition between when the CONFIG_SITE list is generated and
then used here via the cat to create the config.cache and in *all* configure
scripts when reading from CONFIG_SITE.

The race in this case is that the sstate setscene task runs on a package that
contains a site config file (ncurses in this case) and then bitbake decides
that it needs to rebuild and cleans out the site config file, so it existed
for siteinfo_get_files() to find in SITECONFIG_SYSROOTCACHE and then was removed
for the rebuild. When bintuils tried to run the do_configure() task which creates
the binutil's version of config.cache it reads from CONFIG_SITE which now contains
the non-existant site config file. (confused yet ;-)?

Currently the configure script does a test -r to ensure the file is readable
before using it, therefore having the cat succeed regardless of the file being
available is consistent behaviour.

(From OE-Core rev: ffd8f05e5548500199c9b04a174067811ad2c5e7)

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

wget: disable uuid by default

This will ensure a consistent build and not add additional dependency of
util-linux to wget

(From OE-Core rev: 919e4335c5f907fde4b46eb0cd421fbd360ddbae)

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

cross: Clear do_populate_sysroot[stamp-extra-info] so that stamps are not MACHINE specific

Otherwise we will try to run the setscene task once per MACHINE which will lead to
file conflicts in the sysroot for packages like gcc-cross.  The stamps
are already namespaced by TARGET_ARCH which should be sufficient.

(From OE-Core rev: 66d09565783186ab14fa77ecdfe2131a44a265ab)

Signed-off-by: Phil Blundell <philb-mXXj517/zsQ@public.gmane.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

busybox: backport sulogin empty root password fix

This allows system maintenance login if the root password is empty.

(From OE-Core rev: 28bb8fe5c144e02c28bff54b5b81c8da33b9f58b)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemu: Updated from 1.4.0 to 1.4.1

0001-doc-Fix-texinfo-table-markup-in-qemu-options.hx.patch and 3f08ffb4a4741d147634761dc053ed386243a0de.patch were backported in 1.4.1

(From OE-Core rev: d3fd92010a0b965fe1b7696f7c4e6c02d07f85dc)

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

sqlite3: update to 3.7.17

build and runtime tested on armv7a

(From OE-Core rev: c1ad532977f789e6a5e0c1782d58a334f31f3d88)

Signed-off-by: Jack Mitchell <jack.mitchell@dbbroadcast.co.uk>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel: Remove unnecessary cd and rm, fix MODULE_TARBALL_SYMLINK_NAME creation

* DEPLOYDIR is used as target for every other file deployed by do_deploy,
  this was just unnecessary switching CWD to create relative symlink and I
  have no idea why rm -f was used together with ln -sf.

(From OE-Core rev: 8375782a295c6294ea4ab889c309bebf873f3489)

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

patch: diable xattr as default

There is a optional dependency on xattr (and thus the attr package),
disable it by default, to ensure it builds correctly.

(From OE-Core rev: a7de32c8ad1405ed31b620fb99dfe0fe80ac23a9)

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

connman: upgrade to 1.14

- switched to archive releases;

(From OE-Core rev: 2aa82039311a8d84a06ec5fba46c99f64e36e697)

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

perl-tests: convert to ptest

On 05/17/2013 05:31 PM, Björn Stenberg wrote:
> Unless there are clear advantages with patching t/TEST that I have overlooked, I suggest using sed in run-ptest instead.

Ok.. Here is the new commit.

Removed the patch and added a call to sed in run-ptest. See attachment
for the test log.

From: Jesse Zhang <sen.zhang@windriver.com>
Date: Tue, 14 May 2013 02:53:30 -0400
Subject: [PATCH 1/1] perl-tests: convert to ptest

Replace PERL_TEST_DIR with PTEST_PATH, and rename "tests" with "ptest"
in various places. Also add a run-ptest script.

[YOCTO #4292]

(From OE-Core rev: 364cad5d8eecfec74a7be8cf93e75cd63031101f)

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

sudo: set CFLAGS for build mksigname and mksiglist

mksigname and mksiglist are compiled by BUILD_CC for build host. When
there are some options in CFLAGS that BUILD_CC doesn't support,
compilation fails.

Build for arm on a x86 host, if option "-mapcs-frame" is provided, error
occurs with:

| cc1: error: unrecognized command line option "-mapcs-frame"

Pass BUILD_CFLAGS to CFLAGS to fix that kind of failure.

(From OE-Core rev: 6022e81570b80232f272a1aa474e8ced3a089382)

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>

clutter-gst: removed no longer required configure fragment

The current package no longer use DOLT, so the DOLT configure frament is no
longer required.

(From OE-Core rev: b976a332cf24aad18141eafbbe5cd22cfbb91752)

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

clutter: further improvements

 * Removed unnecessary gtk-doc related depends/configure options (handled in
   gtk-doc class),
 * Moved more configure options into PACKAGECONFIG,
 * Fixed missing comma in PACKAGECONFIG[glx] setup,
 * Removed no longer required DOLT-related config fragment.

(From OE-Core rev: a5d9c80fa35e9fd234f14c61b385d9c17f31132b)

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

clutter.bbclass: removed unused function

Clutter and cogl no longer use --with-fpu configure option, so we no longer
require the get_fpu_settings() function.

(From OE-Core rev: ea59330c48cdb7d7106deea3f19c043ba60c23e4)

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

cogl: further improvements

 * Removed unnecessary gtk-doc related depends/configure options (handled in
   gtk-doc class)
 * Moved more configure options into PACKAGECONFIG
 * Fixed typo in RDEPENDS of PACKAGECONFIG[gl] setup
 * Removed no longer needed DOLT-related configure fragment

(From OE-Core rev: 356e00b198608080495641270f4a1634ddc6ff85)

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

runqemu: fix networking issues

After switching from ifconfig to ip, networking stopped working. This
commit contains the following fixes:

 * set a decent broadcast address for the tap device;
 * bring up the device;
 * add the route using ip tool instead of the old route tool;

(From OE-Core rev: a286514e2311f52b54d3571dbac6d34aff39e591)

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>

resolvconf: Update to 1.72

(From OE-Core rev: 6edee60964a1bc367229239175426bdf175efa62)

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

cmake: Update to 2.8.11

Rebased support-oe-qt4-tools-names as some of the changes where
added in the upstream code.

(From OE-Core rev: 13399264e987b698b120688dc5018adb3fa8522d)

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

mc: Update to 4.8.8

(From OE-Core rev: 5779021a7ad251a8359795a6a625e76a6befb3c7)

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

libnl: Update to 3.2.22

(From OE-Core rev: 9376396be8af3dd7603e93127a9f1a0a7ab8ca4a)

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

mklibs: Update to 0.1.37

(From OE-Core rev: 4132b811bff36f79caecb3cfb0b7b0b91c2abfc1)

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

desktop-file-utils: Update to 0.21

(From OE-Core rev: a728571f14e4ee5f97043d8b38f14b5513d8d43d)

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

gnupg: Update to 2.0.20

Removed obsolete patch

(From OE-Core rev: 2e89010cb4a19d525bc43589d487827dd4bc401d)

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

xserver-xorg: enable xvfb

Currently Xvfb has been enabled in openembedded-core. Xvfb is needed
for example in Mauve testsuite.

(From OE-Core rev: 2570d5dcb7bbd50a7b3b8c2345492986be3ed95b)

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

SLiRP support in runqemu

runqemu script now takes argument "slirp" in order to
run networking on the qemu machine, without root privileges.

changed the runqemu-internal script in order not to activate
the tap devices if the option is set.

[YOCTO #1474]

(From OE-Core rev: fa7fd7b1cbcfbd01af1949d2ea09b880a0ae0175)

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

kernel.bbclass: Add dependency on binutils

Compiling the Linux kernel requires binutils; kernel.bbclass uses
INHIBIT_DEFAULT_DEPS so it had better depend on binutils explicitly.

(The lack of this dependency isn't always a problem because binutils
is required to build gcc-cross but if gcc-cross is reconstructed from
the sstate cache then gcc-cross's dependency on binutils-cross is
ignored due to being in the safe dependency list in
setscene_depvalid.)

(From OE-Core rev: 37beb7bdab78de5253a894f35afafa34c13a00f5)

Signed-off-by: Mike Crowe <mac@mcrowe.com>
Acked-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glib-networking: update to 2.36.2

(From OE-Core rev: 5cf2fa6cb4b18d1a436b361b805822f7ed391d80)

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

glib-2.0: update to 2.36.2

This patch was build- (gcc 4.7.2/4.8.0) and run-tested with my standard
xfce-/gnome2-images

(From OE-Core rev: f4f5d41f6cd262379daa8a00699a64f0df6fd9e7)

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

librsvg: add sstate dependencies so that the SVG loader works at staging time

Add additional dependencies so that when the sysroot is populated the gdk-pixbuf
SVG module can be loaded.

(From OE-Core rev: 2b68b08f47baf2fc2fa896dde0cc66297441b6b3)

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>

librsvg: small cleanups

Re-order fields to a more logical order.

Remove the unrecognised --disable-mozilla-plugin option.

Remove the unrequired setting of GDK_PIXBUF_QUERYLOADERS, it's found
automatically.

(From OE-Core rev: a209b3858c6f0c45accf2b131b734d44230472ca)

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>

librsvg: clean up native depends

inherit gnome was only used for the SRC_URI, so just inherit gnomebase and
remove the native-specific DEPENDS which was only required as the gnome class
pulls in an impossible hicolor-icon-theme-native dependency.

(From OE-Core rev: dc1fbdd925d4dde294c4b63466e67084ce1ace89)

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>

librsvg: add more PACKAGECONFIG options

Add options for the gdk-pixbuf loaders and the GTK+ 2 theme engine.  The theme
engine is generally unused so don't enable that by default, but enable the
gdk-pixbuf loader and also enable the croco feature which is required for
parsing CSS embedded into SVG.

(From OE-Core rev: 228d15b40500ecfb182078171e33ffa28373024f)

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>

gdk-pixbuf: remove native loader cache generation, use the class

We assumed that the only relevant gdk-pixbuf loaders in the native environment
were the ones in gdk-pixbuf itself, as the icon cache is only for PNG files.
However, glib-compile-resources can transform SVG files to raw image data, and
done natively this requires the SVG loader to be registered.  The current
implementation relies on this assumption by generating the cache based on the
staging directory during install, so if gdk-pixbuf-native is re-installed to the
sysroot after librsvg-native it will overwrite the loader cache.

So, remove the code in do_install that updated the cache, and rely on the new
logic in pixbufcache.bbclass that updates the cache when it's installed into the
sysroot itself.

(From OE-Core rev: 3a7625c23b3fbd163dcd4036767b194438ec238d)

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>

pixbufcache: update the loader cache when installing natively

Register a sstate postinst function so that when installing a native package,
the gdk-pixbuf loader cache is updated.

(From OE-Core rev: 7cbde5d8f2b88ec4e79cb0d564ee0f3c9baa7c2d)

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>

util-linux: Update to 2.23

Remove license patch as it is integrated upstream.
Add backports of upstream loopdev regression fixes.
Updated uclibc-__progname-conflict.patch because it didn't apply.
Added bash-completion and partx sub-packages.

(From OE-Core rev: 54629315502247c5751c351b5792838f86dd1ea8)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: cooker/cookerdata: Fix up event handling for hob

The init function call was clobbering data hob relies on injecting, fix
this so the data injected by hob isn't overwritten.

(Bitbake rev: 64ca1aa1fd5f0473196b6cd805637fa644a92173)

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

bitbake: build: Avoid empty promises

The "see xxx for further information" is misleading since it is just
the same information. Clarify just to mention the that this is the location
of the logfile without any empty promise.

[YOCTO #4343]

(Bitbake rev: 7088c0e8553dd3c408b5bc06f8c34d5b72e9ea9a)

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

bitbake: bitbake-layers: fix show-appends and flatten for recent cooker changes

Fixes the show-appends and flatten subcommands for recent refactoring in
cooker.

[YOCTO #4536]
[YOCTO #4535]

(Bitbake rev: 54817d0de667941ee68fe3490684159e2d90e9c5)

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

bitbake: cooker/cookerdata: Improve configuration object handling

Originally it seemed like a good idea to keep the parameters around. Having
seen this in real life use, its incorrect, we should pull all the data we need
into the cooker's configuguration and then use this to build the datastore.

Being able to just build the datastore from the parameters seemed like a good
idea but having a dummy cooker configuration object is now looking like
the better option.

This also fixes failures in hob since the parseFiles command can call
into cooker directly now and reset the configuration prefiles and postfiles
at will, rather than the indirect calls before which were breaking the datastore
(e.g. BBPATH wasn't set).

The cleanup this allows in tinfoil illustrates how this change makes more sense.

(Bitbake rev: f50df5b891bf318f12fc61c74adfcc626cc6f836)

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

bitbake: fetch2: Fix missing assignment

The assignment to True was missing from the code, well spotted Saul!

(Bitbake rev: e493fe8cb4953935f01361ffc0240e5818ebb283)

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

bitbake: cooker/cookerdata/event: Improve class handlers management

Similarly to the execution context changes, establish better lifetime
management API of the class event handlers.

(Bitbake rev: 54e35a6cceead9521f8b1dacd48e55064e85c8bd)

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

bitbake: cooker/cookerdata/utils: Improve context management

The current execution context management for bitbake is ugly and the
use of a global variable is nasty. Fixing that is hard, however we
can improve things to start to establish an API for accessing
and changing that context.

This patch also adds in an explicit reset of the context when we reparse
the configuration data which starts to improve the lifecycle of the data
in setups like hob.

(Bitbake rev: 6c3281a140125337fc75783973485e16785d05a1)

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

bitbake: event: Drop Handled/NotHandled status return values

These have long since been deprecated, lets remove them.

(Bitbake rev: 3dc83bbb1bf387bb7ecea2e17f0f72cfccecba92)

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

bitbake: cooker.py: Remove global scope class event handlers

Its confusing when an event handler in for example a bbclass receives
events for a recipe which doesn't use the class. This is due to the
class event handlers having confusing scope. Worse, the issue is not
deterministic and the events a handler will see depends on the parse
order.

To avoid these issues, remove the class event handler global scope
and only have class handlers recieve events for recipes using the
class.

(Bitbake rev: 7fb95f3d133e440d463d2faa7151c731f8e1ae96)

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

bitbake: methodpool: Retire it, remove global method scope

Having a global method scope confuses users and with the introduction
of parallel parsing, its not even possible to correctly detect conflicting
functions. Rather than try and fix that, its simpler to retire the global
method scope and restrict functions to those locations they're defined
within. This is more what users actually expect too.

If we remove the global function scope, the need for methodpool is reduced
to the point we may as well retire it. There is some small loss of caching
of parsed functions but timing measurements so the impact to be neglibile
in the overall parsing time.

(Bitbake rev: 4d50690489ee8dc329a9b0c7bc4ceb29b71e95e9)

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

bitbake: methodpool: Conflicting methodnames should be a fatal error

When this error occurs, the build should stop, not continue uninterrupted.

[YOCTO #4460]

(Bitbake rev: acd6d7ffa8813b3b11cad9145e8e614a695ae04a)

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

sanity.bbclass: Drop dubious print

(From OE-Core rev: be4cd4614a2a702130d18e9d082fd83a15faa013)

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

buildhistory: Sync with bitbake changes

Adjust to the recent changes in the git fetcher. This code should work
with current and previous versions of bitbake and can be cleaned up when
we move to new bitbake versions.

(From OE-Core rev: 835b64d4e9ed7b627b56a75d529f1a403c5db1a8)

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

rpm-postinsts.bb: enable postinst logging

Enable postinst logging by checking the configuration in ${sysconfdir}
/default/postinst.

In this way, the postinst logging is enabled if 'debug-tweaks' is
in IMAGE_FEATURES, and at the same time, we avoid unnecessary rebuild
if IMAGE_FEATURES is changed.

[YOCTO #4262]

(From OE-Core rev: 6f2aa32f10c24c84e581128bb3a976ef071197ac)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

opkg: modify the run-postinst script to enable postinst logging

Enable postinst logging by checking the configuration in ${sysconfdir}
/default/postinst.

In this way, the postinst logging is enabled if 'debug-tweaks' is
in IMAGE_FEATURES, and at the same time, we avoid unnecessary rebuilt
if IMAGE_FEATURES is changed.

[YOCTO #4262]

(From OE-Core rev: 35be6ffc19a5156aa029397707f1e6869684b821)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dpkg: modify the run-postinst script to enable postinst logging

Enable postinst logging by checking the configuration in ${sysconfdir}
/default/postinst.

In this way, the postinst logging is enabled if 'debug-tweaks' is
in IMAGE_FEATURES, and at the same time, we avoid unnecessary rebuilt
if IMAGE_FEATURES is changed.

[YOCTO #4262]

(From OE-Core rev: 881880992ac0edc5f928f7e3d2a8f3f993284df6)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image.bbclass: add postinst_enable_logging

Add a function postinst_enable_logging, so that when 'debug-tweaks'
is in IMAGE_FEATURES, we create ${sysconfdir}/default/postinst config
file, which is sourced by run-postinst scripts to determine whether
to log or not, and where to log.

[YOCTO #4262]

(From OE-Core rev: 99175cabc3936733dd92fac5ebc6f865b864fe92)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scripts/bitbake: Handle the case where git isn't installed cleanly

Currently the user sees ugly errors if git isn't installed, this patch
cleans up the code to correctly handle that case.

(From OE-Core rev: aeb704fee8b4ffeaeddcdb36ae4e1d62c264ce42)

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

task.bbclass: Drop deprecated class

Any users have had time to adapt to this change by now, drop the old class.

(From OE-Core rev: 9b64b68c93c71c503ef26fa440b974b82438dc88)

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

update-alternatives: Drop deprecated code, update to show error message

This code has been deprecated for a while and confuses the class, lets drop it
and just give the user hard error messages instead of the current warnings/fixups.

(From OE-Core rev: 72579e1fe49e8bc66c9f5850a2c679ce8941c85d)

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