scripts/qemuimage-testlib: Output a slightly better error if expect is missing

(From OE-Core rev: bfbbb420c0a6b315f73d3d5e463a518098ca6ef0)

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

udev-extraconf: blacklist /dev/md

Do not mount /dev/md by default via udev, this resolved a problem
with the sanity test failing due to seeing the error while attempting
to mount /dev/md0

(From OE-Core rev: 07a2825c6f4ad3e5e3970cd1a89233bd795c68cf)

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

Distro_tracking: Update Manual Check Date

Checked the following Upstreams:
apt
libaio
sysstat
unzip
zip
linexif
boost
libcheck
ncurses
sysfsutils
util-linux - Offline
lsb
eds-tools
libmad
lame
glew
squashfs-tools

(From OE-Core rev: 8665e4a47e944b53043a003eda12a447daee51c5)

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

libarchive: Remove obsolete comment and empty line at the end

This is a fix up for

        commit fb19df5b21e551c5dfdfa340438952560c5fa528
        Author: Xiaofeng Yan <xiaofeng.yan@windriver.com>
        Date:   Mon Nov 7 20:03:53 2011 +0800

            libarchive: update to 2.8.5

            Remove patch "0003-Patch-from-upstream-rev-2516.patch" because it has been merged
            to source codes.

            (From OE-Core rev: 1df17fb4c22e00e6a180777dd23efb6d5cb1590a)

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
removing a now obsolete comment because the undistributable content was removed [1] from upstream’s tarball.

Also remove an empty line at the end introduced in the above commit.

[1] http://code.google.com/p/libarchive/issues/detail?id=162

Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Introduce new SERIAL_CONSOLES to add multiple consoles for your MACHINE

Just define additional serial consoles like so:

SERIAL_CONSOLES="115200;ttyS0 115200;ttyS1 ... 115200;ttySN"

Also be sure to remove SERIAL_CONSOLE (lacking the S) from your
machine as they can conflict.

(From OE-Core rev: a61b5e980d45df4dc2c83dd13996b0563fddf803)

Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

data_smart: Add appendVar/prependVar functions

This patch adds appendVar and prependVar functions to the data store
meaning python code would no longer have to do the getVar, append and
the setVar dance that much of the current python code does.

It also adds corresponding variants for flags.

Currently there is no spacing added by these functions. That could be
added as a parameter if desired.

If these functions turn out to be hotspots in the code, there are tricks
that could potentially be used to increase the speed of these specific
operations within the datastore.

(Bitbake rev: 4a4046268f84b85559eea2c4b6a6004ad8cccb77)

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

Add FAKEROOTNOENV variable

Add a FAKEROOTNOENV which does the opposite of the FAKEROOTENV variable
and is data loaded into the environment for tasks without the fakeroot
flag.

The intent here is to provide a way to control the environment when we
aren't needing fakeroot context which allows us to unload the preload
from memory entirely and gain a valuable speedup.

I'm not 100% happy with needing this at the bitbake level, particularly
with the cache hit but it does give a valuable speedup.

(Bitbake rev: 05c29ab0b2ad3c521414cabb6a92bca15c6e919c)

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

cooker.py: Ensure only one copy of bitbake executes at once

The bitbake codebase makes assumptions that only one copy is active
against a given build directory at a given time. This patch adds a
lockfile in TOPDIR to ensure that is the case.

Note that no unlock is needed, that is automatically dropped when
execution terminates.

(Bitbake rev: 6004cbf36c980c2574c1c9153df81a7c55317e17)

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

utils.py: Fix lockfile retry handling

The lockfile retry parameter is expected to return immediately after
attempting to take the lock. There was a bug in the logic which this
patch fixed to ensure it does that.

(Bitbake rev: f421ef819f00ac659504d9af41bcc8323422ff8c)

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

codeparser: silence non-literal warnings for vardeps

If the vardeps flag is not None, we now silence the warnings about
non-literal usage for that variable.

(Bitbake rev: e724b9f417d1baf898f5afc6376c73c1a2ad8db9)

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

codeparser: drop expand tracking

There are two usual cases involving bb.data.expand:

- Calling it with a string literal -- "bb.data.expand('${FOO}/${BAZ}/bleh', d)".
- Calling it on getVar results (legacy) -- "bb.data.expand(bb.data.getVar('FOO', d), d)"

Nothing in any of the usual layers uses it in any other way, and I'm
having trouble coming up with any real use cases beyond this. The first
of the above cases is already tracked, via the expandWithRefs called
on the python code string. The second didn't emit a warning anyway,
since the getVar was already handled.

Given this, I see no reason for us to maintain explicit expansion
tracking. Further, we weren't using its results anyway (the var_expands
member).

(Bitbake rev: 405dfe69e6a608826e599ebf2f83ef8cf5083b96)

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

codeparser: accept a name for better messages

- If a name is passed to the parser, prepend the messages with "while
  parsing <name>:". This gives a bit more context.
- Tweak the warning messages slightly (they had to be altered anyway to
  inject the variable being parsed).

Before:
  DEBUG: Warning: in call to 'bb.data.getVar': argument ''%s' % var' is \
         not a literal

After:
  DEBUG: while parsing emit_pkgdata, in call of bb.data.getVar, argument \
         ''%s' % var' is not a string literal

(Bitbake rev: 1060193ae4d54e667735dbff5d1d2be49a3f95c9)

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

codeparser: simplify how we compare the called node names

With the previous method, using the compare_name methods, we split the
requested match name by '.', reversed it, then compared them piecemeal
during the node traversal. The new method walks the nodes and hands back
the name of what's being called, and then we check that. This also
consolidates the two different implementations of traversal of the
attribute/name nodes (one in compare_name, one for the execs).

(Bitbake rev: 84e535b5165c7e936c5b1486bdf4626ed3649f5f)

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

codeparser: merge the nested python parsing classes

The split is even less necessary now that we use ast.walk rather than an
actual NodeVisitor subclass.

(Bitbake rev: d6c44fac184abae8395bfa7078f06675218aa534)

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

useradd: Add missing DEPEND on shadow

Without this rootfs generation fails as an RDEPENDS is added
but the package might not have bneen built.

(From OE-Core rev: bfe70c6446e6686f826f01040ba74c7d7d28bf42)

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

avahi: remove USERADDPN

USERADDPN is no longer used; remove it.

(From OE-Core rev: ed7e7a8e4d00cd45c74dc233c8b574d3978755d8)

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

useradd-example.bb: update example documentation comments

Clarify that only packages listed in USERADD_PACKAGES will
include the user/group creation code.

(From OE-Core rev: 70aaac37968bf2b35d6a536c3f3f69fe3620255c)

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

useradd.bbclass: do not modify -nativesdk packages

Exclude the addition of user/group code and RDEPENDS changes for
-nativesdk packages.

(From OE-Core rev: 2f057dd905ccb497890ce73ac4e4c256edcf0351)

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

useradd.bbclass: only modify packages in USERADD_PACKAGES

Previously we injected the user/group preinstall script into all
output packages. This fixes that so that only packages listed in
USERADD_PACKAGES get modified.

It also removes the USERADDPN variable, which is no longer needed.

(From OE-Core rev: 2f73466eb5018040a123ccb0e2af8c519525f958)

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gconf: Drop polkit dependency

(From OE-Core rev: 11506f702a83a77d06d259ce05c89808d53a9661)

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

gconf: add upstream GNOME gconf 3.2.3 and drop gconf-dbus

The D-Bus backend has been integrated into upstream GConf so we can
switch to upstream and drop gconf-dbus.

I've gone for a release in the 3.2 series as we disable Gtk+, and
therefore are not impacted by the gtk+3 changes, and the D-Bus backend
was unstable before this release.

(From OE-Core rev: f952b1bc8a4f307e2e2f3941f50becd72d88421d)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mime.bbclass: No need to import os.path, it then breaks the os module

(From OE-Core rev: b4457a9605988feecadfd49df22ef57245c4cd86)

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

classes/[gnome|gnomebase|mime]: enhance gnome related classes

This patch pulls in the gnome related classes from oe-core which
adds extra packaging rules and functionality whilst modularising things
so that one can get a subset of gnome functionality without adding a lot
of extra dependencies.

These aren't an exact copy of the classes from meta-openembedded, notable
differences are:
* gnome.bbclass - I dropped the BBCLASSEXTEND
* mime.bbclass:
  - updated coding style
  - use which to find update-mime-database program rather than hard coded
  - fix typo in populate_packges_append such that it's actually called

CC: Koen Kooi <koen@dominion.thruhere.net>
CC: Richard Purdie <richard.purdie@linuxfoundation.org>
(From OE-Core rev: 6ead00eba265211eae5a2423c155403c8c6d54d0)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

shared-mime-info: package runtime data separately

The freedesktop.org.xml is required when updating the mime database but
otherwise not much use. Therefore package it separately, rather than
removing it completely, so that it can be used as required for on-device
updates without adding almost 2MB to every image which hase shared-mime-info.

(From OE-Core rev: 07ee2d7e6b2f571fe405464838e376cf03e5e8e6)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

busybox: add grep to temporary links during uninstall

In the busybox package prerm we set up some temporary links and modify
PATH so that certain utilities are provided for the purpose of running
update-alternatives; if grep is not among these then you get errors when
removing busybox, so add a temporary link for grep as well.

(From OE-Core rev: 013eca09c863862cc6b7ee3bc22923bf8fb42956)

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

classes/package_rpm: disable uninstall scripts for upgrades

Our current assumption (based on the behaviour of opkg) when writing
recipes is that prerm and postrm do not get called during an upgrade.
When using rpm however, these are mapped to the rpm "preun" and "postun"
events which occur after postinst for upgrades, and when these contain
removal type operations (such as update-alternatives --remove) this
causes problems.

This patch wraps each preun and postun script for rpm in a check that
determines whether or not the script is being called during an upgrade,
and skips the entire script if it is, which mimics the behaviour of opkg
under the same conditions.

Fixes [YOCTO #1760]

(From OE-Core rev: 1d3f37dc9a43ba6d6beb7b4530c077f239032b99)

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

python: skip setup.py 'import check' when cross-compiling

build_extension() in setup.py, as part of the build process, does an
'import check' on the built extension.  The import check in turn
dlopen()'s the shared library associated with the extension, which
isn't something that makes sense if that library was cross-compiled
for a different architecture.

This was noticed with an x86_64 target that was compiled with avx
support, because it caused 'illegal instruction' exceptions:

| /bin/sh: line 1: 14575 Illegal instruction ... -E ./setup.py -q build

For other target architectures, it doesn't necessarily cause illegal
instruction exceptions, but still fails.  For example, on arm, the
failure pathway causes this warning:

*** WARNING: renaming "cmath" since importing it failed: .../cmath.so:
    wrong ELF class: ELFCLASS32

This patch to setup.py and the associated recipe changes allow the
whole 'import check' logic to be skipped when cross-compiling.

(From OE-Core rev: 25fae81538a92e15eab3fc169ebce44505f67839)

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

libzypp: fix mishandling of hyphenated arches

Several hyphen-to-underscore translations were missing, causing
compiler errors trying to build arches with hyphens in their names.
This adds the missing translations.

(From OE-Core rev: 5be9785f344ec4d7580f7ec68e29dba9fceb0a0a)

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

gmp_5.0.2: Set CC_FOR_BUILD to BUILD_CC

CC_FOR_BUILD was compiling the test programs using the target's
compile options and executing those on the host, causing errors such
as:

/bin/sh: line 1: 15032 Illegal instruction     ./gen-bases table 64 0 > mpn/mp_bases.c
/bin/sh: line 1: 15033 Illegal instruction     ./gen-bases header 64 0 > mp_bases.h

Export CC_FOR_BUILD using BUILD_CC to fix the problem.

(From OE-Core rev: 68cca5ca15cbdd53748ec130fb6f20cbb3fb5072)

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

Add new IMAGE_CLASSES variable for classes for image generation

Allows us to import classes only for images and not to the global
namespace

(From OE-Core rev: 49dcb301ab39327554d86d23cf6f8d435d7a7351)

Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

distro_tracking_fields.inc: Modify the status of qmmp

update qmmp to 0.5.2

(From OE-Core rev: 1f8b688bf8bddaacf90cc034e65bc95cf4310bcb)

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qmmp: update to 0.5.2

qmmp: remove old recipe and add new recipe

Use do_split_package for each grouping and name them correctly besides updating package.
The format to name plugins of qmmp:
qmmp-plugin-<group>-<library>   with the Description: Qmmp<Group>  plugin for %s"

(From OE-Core rev: fff02c7c2c1c6ab0ceb1f888134d381d8b612884)

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
Fix dESCRIPTION -> DESCRIPTION

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

distro_tracking_fields.inc: update to libarchive to 2.8.5

Change the status of libarchive.

(From OE-Core rev: 3326df3ba8bcf848322c2a7b3782aa204825200e)

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libarchive: update to 2.8.5

Remove patch "0003-Patch-from-upstream-rev-2516.patch" because it has been merged
to source codes.

(From OE-Core rev: fb19df5b21e551c5dfdfa340438952560c5fa528)

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dhcp: move server configuration to dhcp-server-config

(From OE-Core rev: ca183f84c2f3e132059e2f63d7ed8ca27987a99e)

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

dhcp: rename dhcp4.inc to dhcp.inc

(From OE-Core rev: af51496a9a35ecee0487d554f0493c662c335228)

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

dhcp: drop unused dhcp3.inc file

(From OE-Core rev: 625ca73a16f6ef1da7a5c42450af715e41172e52)

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

files/device_table-minimal.txt: add /dev/kmsg

(From OE-Core rev: 1a340471694204937981513dee4cc24bc2dc6f7e)

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

libcap: fix sstate for native package

The 'lib' option needs to be given on target and native builds
otherwise it installs the binaries at ${libdir}64 when host is 64bit.

(From OE-Core rev: f768ef66c107410d4e81a69543d41910bbc6a26e)

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

pseudo: Uprev pseudo to version 1.2

This adds a new feature, PSEUDO_UNLOAD, which can be used to eliminate
overhead of LD_PRELOAD when no longer necessary.

Also the, clone(2), support on Linux has been updated to resolve some
potential defects in the previous implementation.

(From OE-Core rev: 77fe9dd8fa0393132ac6aba00d5659c6781fbbde)

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

clutter-gst: fix patch filename

Munged the patch filename somehow. Apologies for the noise.

(From OE-Core rev: 157db263b50fabf4afcbea3c2a5364c0752ae0e2)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

useradd.bbclass: handle nativesdk case

* without this patch, building dbus-nativesdk leads to a missing
dependency on 'base-passwd-nativesdk'
This was added by commit 46e6c3fa8034b12d178d605f3f5d7efe69671a13
* this patch handle the nativesdk case in the class useradd
* close bug 1702 http://bugzilla.pokylinux.org/show_bug.cgi?id=1702
* v2 from Scott Garman with Richard Purdie's tricks

(From OE-Core rev: 140a3507fb5c14cd9bcebe4304f491aa1c5c47a2)

Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Scott Garman <scott.a.garman@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

multilib: Drop MULTILIB_IMAGE_INSTALL

There should just be a single IMAGE_INSTALL variable. If the package
backends need this split into different multilib components they should
be responsible for doing this, not the user.

This commit removes the MULTILIB_IMAGE_INSTALL variable.

[YOCTO #1564]

(From OE-Core rev: 7736862a74c92fe1afe42e170822be13117575c2)

Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

contacts: fix packaging of icons

(From OE-Core rev: 1f4028337d5e288e239f44ef34e1d707b785273e)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio: make X11 dependencies optional and add gtk+

Use the x11 distro feature to determine whether --with-x should be
passed. Further, if we're using X11 add gtk+ to DEPENDS (it's required
to build).

(From OE-Core rev: a37ed5b42717463cf1a836fd011132b16938765f)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

connman: split scripts into separate package

The connman scripts are useless without a Python interpreter so split
them into a separate package and add Python to its RDEPENDS.

(From OE-Core rev: 281510acd0ee3221dd804d7d9de567be2be675f4)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gypsy: fix packaging

(From OE-Core rev: 3c272ce9df811281029d028e96ab6bc644645592)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libcanberra: add libvorbis to DEPENDS

(From OE-Core rev: 531151fdeba3779ba6f0976fc08aa8da483600f7)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mx: add mx toolkit 1.3.2

A widget toolkit built on Clutter

(From OE-Core rev: 67b2060714c70cf39224e5fbe83ca36263595c54)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

clutter-gtk-1.8: add clutter-gtk 0.11.4

(From OE-Core rev: e5568f9a569e6ff1cf45553edb18923a4f7843e9)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

clutter-gst-1.8: add 1.4.2 for use with clutter 1.8

(From OE-Core rev: acca04830aad1368f0a5cc2f48302cf272731174)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

clutter-1.8: add 1.8.0

(From OE-Core rev: d59578fd9e2c45bdc2f030d68c84d05c5ad5be6b)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cogl: add cogl 1.8.0 recipe

Cogl has been split into a separate recipe as of the clutter-1.8 series

(From OE-Core rev: 3bddb7cd3400857ac724e91cb5b663bcf386ab37)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dbus: fix install for virtclass-nativesdk

* 46e6c3fa8034b12d178d605f3f5d7efe69671a13 changed do_install
which now fails for nativesdk (chown messagebus leads to no
such user)
* tested by building meta-toolchain-qte and running the generated
sdk

(From OE-Core rev: 5818a885df489f4bc9579d17c6b0efa7777f5ccc)

Signed-off-by: Eric Bénard <eric@eukrea.com>
Acked-by: Scott Garman <scott.a.garman@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

useradd.bbclass: fix how RDEPENDS is setup

Fix bug where only packages named PN included base-passwd in
RDEPENDS.

This fixes [YOCTO #1727]

(From OE-Core rev: 2c55d51afd71d708a54afc8377e10c4f80f810e3)

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lighttpd 1.4.29: rename index.html to avoid clashes with DL_DIR

As Martin Jansa pointed out before, bitbake will silently peek in DL_DIR before FILESPATH and use files from there if found.

The failure mode for lighttpd involves a 404 redirect placing index.html into DL_DIR, which will end up in the lighttpd packages. In my specific case iproute2 hit the linuxfoundation 404 redirect so lighttpd.ipk now serves the linuxfoundation frontpage :)

(From OE-Core rev: 0809f98e995d3a2ce24ad25bade3dcd5154061fd)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemugl: switch to new git repo

The qemugl git repository is now on yoctoproject.org.

(From OE-Core rev: da1759308bd2d8ca3bf4dc077c164aadb5ce3916)

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

core-image-minimal-initramfs: force IMAGE_FSTYPES

If the user has set their own value for IMAGE_FSTYPES, they may have
disabled the cpio.gz image type, preventing the initramfs from being
produced in the format that image-live.bbclass expects; so force
IMAGE_FSTYPES to cpio.gz within the initramfs image recipe.

(From OE-Core rev: 571919f49cbe6f8f5b21de844b00761314e697fd)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Prepend cpio.gz instead

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

allarch.bbclass: disable shlib and debug symbol processing

all-arch packages should not contain any binaries that need processing,
so disable the shared library dependency processing and debug symbol
splitting/stripping to save some time during packaging.

(From OE-Core rev: 9534b3efe94fb9a9cd927f29c375bbdb74b84035)

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

fotowall: disable videocapture since we do not have a webcam

(From OE-Core rev: 837a0cf30491fefb051c036f0cdc0ac661615a9b)

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

linux-yocto_3.0: update base to v3.0.8 + meta config changes

The board branches are updated to have v3.0.8 as their base kernel
revision, and the version in the recipe is updated to match.

This commit also updates the meta branch to pick up the following
4 commits:

  4095bb5 meta/common-pc-64: remove igb
  17565fa meta/crownbay: use power/intel feature
  8948937 meta/sugarbay: use power/intel feature
  fcbc8a9 meta: add power feature

(From OE-Core rev: c237410bbc177495dab2b12b3a7aa4a75fb61efa)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libnl-2.0: add PE/PR bump for upgradable patch for meta-openembedded users

(From OE-Core rev: 2260b18590416940eec26aaf3d68e510ceff8d31)

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

libnl-2.0: split to more packages, as meta-openembedded does

(From OE-Core rev: 8720e063c7b43c278b3bb406b45390ed03f8ac96)

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

libnl-2.0: add patch from meta-openembedded to fix pkg-config file

(From OE-Core rev: 72227178bc74d6e2e24f8df6176c3d45b640e860)

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

libnl-2.0: move fix-pktloc_syntax_h-race.patch to libnl-2.0 subdirectory and merge with fix-makefile.patch

(From OE-Core rev: a4882cd6f98c5b3df80ba96536d94d9f556f77a2)

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

iproute2: update to 3.0.0 to fix build with updated iptables

(From OE-Core rev: addcb5106ca0b1849299d6ad4e734135da29b0ab)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

connman 0.75: bump PR for iptables library change

connmand[180]: /usr/sbin/connmand: error while loading shared libraries: libxtables.so.5: cannot open shared object file: No such file or directory

$ dpkg-deb -c ipk/armv7a/iptables_1.4.12.1-r0_armv7a.ipk  | grep libxtables
-rwxr-xr-x root/root     35252 2011-11-01 20:16 ./usr/lib/libxtables.so.7.0.0
lrwxrwxrwx root/root         0 2011-11-01 20:16 ./usr/lib/libxtables.so.7 -> libxtables.so.7.0.0

(From OE-Core rev: 6c7b71d9a3e855760acacef6949665a305bac2c0)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python: improve packaging

* move 2to3 to separate package and include lib2to3 (was in python-misc)
* fix pattern for python-unittest (was in python-misc because it's in subdirectory now)
* add pydoc_data to python-pydoc (was in python-misc)
* add more stuff to smtpd, audio, codecs, ctypes, html, io, json, mime,
  pickle, stringold, xmlrpc
* move all FILES_ details from python recipe to manifest generator so it's in one place
* added manual line break in FILES_${PN}-core, because git send-email
  doesn't like too long lines
  $ git send-email -1 dfaae65839f0ab23e5b2ae2a68df0f370bca84d2
  fatal: /tmp/k8zbDajUNP/0001-python-improve-packaging.patch: 64: patch contains a line longer than 998 characters

(From OE-Core rev: f17f6b28ed2f62250f8690617e9126a43c3a8020)

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

distro_tracking_fields: updates for sudo, mtools, grep, and openssh

(From OE-Core rev: 696c26c3244cb11b1afe1870e54c838ea7815272)

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image_types bbclass: use 4k bytes per inode so we don't run out of space immediately

genext2fs only creates the minimum number of inodes, after this patch it will scale with the rootfs size

(From OE-Core rev: c31cb0bdc5a61d2d9f21a2cea34c3d8ac3b47cb9)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libcense.bbclass: fix OpenSSL mapping

[YOCTO #1712]

(From OE-Core rev: 56799ebcb5c55a7fc75458fc2be2e69a67e8fd21)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Fixed YOCTO bug format and location

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

tslib: fix the bug with loading libts-1.0.so

Touchpad did not work in the qtdemoE if the library libts-1.0.so was not loaded
manually using the LD_PRELOAD variable. This problem was fixed in the tslib mainline
https://github.com/kergoth/tslib after the 1.0 release. We just import the patch.

(From OE-Core rev: 0ba6d91dc527908740890c896b834e7216b0d2fb)

Signed-off-by: Dmitry Cherukhin <dima_ch@emcraft.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev-164: Update init script to do an explicit add action

With udev 152 or greater the default action for 'udevadm trigger' was
modified to be 'change' instead of 'add.

To ensure initial coldplug events at boot are seen be scripts the are
expecting them as 'add' events we invoke udevadm with an explicit
'--action=add'.

(From OE-Core rev: eacafd21999ab37b60af29dc3e626c441716ef66)

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

buildstats: Fix for buildstats on tmpfs

tmpfs/encryptfs/(and most likely, but not confirmed)ramfs TMPDIRs
cause diskstats to choke. No device entry ends up in /proc/diskstats
for these fs types, which ends up causing the failure.

The short term solution is to exclude these fs types from diskstat
collection. Longer term we will want to see if we can collect
meaningful diskio for each of these, and other, use cases, but for
this cleans up Bug 1700.

[YOCTO #1700]

(From OE-Core rev: 2b14046c12855b6f484ba5bd6bc0a8022de6873e)

Signed-off-by: Elizabeth Flanagan <elizabeth.flanagan@intel.com>
Corrected YOCTO bug location and format

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

dtc: Add patch to correctly install shared libraries and links

(From OE-Core rev: 4c4143e0f0382118448c7c2a452dafac0db5130f)

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

dtc: update to latest git version

Also remove patches that are no longer needed, as well as use the
installation that is provided by the Makefile. Put all the extra
stuff in a new dtc-misc package

(From OE-Core rev: 43213fd610da7b550e6d738e224b49004776f799)

Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cogl: COPYING file Updated FSF Address

(From OE-Core rev: f3a7b6aaa9be4763d13214c3264b86a1c7bf5e9f)

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

scripts/combo-layer: skip empty commits

If a commit is empty (for example, commits brought over from svn where
only properties were changed) then attempting to apply it with "git am"
will result in the error "Patch format detection failed", so skip it
instead.

(From OE-Core rev: 5754bc7bcc2f57b6b56e67e2900eeaab5bc184d8)

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

scripts/combo-layer: improve some messages

Fix some grammar.

(From OE-Core rev: cc34f1bf136409d2e31c443dd5334c3a35d208af)

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

scripts/combo-layer: auto-commit updated config file

If the config file is tracked within the combo-layer repository and it
is updated at the end of the "update" operation (because last_revision
has been changed), then automatically commit the file. This ensures that
multiple people can perform updates on different machines without the
last revision information going missing.

(If the file is outside the repository or is masked via .gitignore, this
will do nothing.)

(From OE-Core rev: 14f89acb43b0d9bec09b94668fbb056c7b5b1456)

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

scripts/combo-layer: fix dirty repo check

Fix the dirty repository check to detect any changes, including
untracked files and uncommitted changes in the index.

(From OE-Core rev: f03fa8bbafa82ce6b09c08bf86011b6c51eb3531)

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

scripts/combo-layer: make component repo branch configurable

Add an optional per-component branch setting to allow specifying the
branch instead of always using master.

(From OE-Core rev: 8e2b8b05607103acd539808c5ab0cc80c0d481fc)

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

scripts/combo-layer: tidy up config example

Move all example configuration to the example config file, tidy up a few
long lines and fix some grammar.

(From OE-Core rev: ec099a32243ebc7eecd86e4bf40ed38da4af3fe5)

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

scripts/combo-layer: make init set up initial component data

The "init" subcommand will now copy the current state of each component
repository into the combo layer repository, set last_revision for each
component, and then use "git add" to set up the initial commit (but will
not actually make the initial commit - that is left up to the user).
Also take the opportunity to bump the version number.

(From OE-Core rev: d4d491aa118b6cbe895167c62a46d883abc7b4ab)

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

scripts/combo-layer: ignore merge commits

The combo-layer tool had trouble handling merge commits as they were
included in the revision list but not the patches; these are now
excluded from the revision list. Note however that this will not handle
merge commits that resolved a conflict; since "git format-patch" cannot
construct a linear change history over such merges there is nothing we
can currently do with these.

(From OE-Core rev: b6b6b215675631d0e505ffc8b4890f85fb3169f9)

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

bitbake.conf: Drop remaining TERMCMD pieces and document OE_TERMINAL usage instead

(From OE-Core rev: ddb825c87dd3e59d50841a993080a00d1459b1e7)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
[Yes, this one is against meta-yocto, I'll fix it to apply the same
change to OE-Core's local.conf.sample in the final version]

patch: Convert to use oe_terminal

Unfortunately we can't access oe_terminal directly from patch.py
so we have to pass in the correct terminal function pointer.

[YOCTO #1587]

(From OE-Core rev: 9e0a21dda24f285a1c4878488e887485a749f3f2)

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

bash: Ensure we fully reautoconf the recipes so site data is used

This ensures bug 487 (missing job control functionality) really gets fixed.

[YOCTO #487]

(From OE-Core rev: 08b78066bd5a9ff2819a42eb4263ee0a78cddb97)

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

rt-tests: update to 0.83

Fetch from github, kernel.org uri is gone

(From OE-Core rev: e3c003282afb93ec52882496400b042620ab00ef)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rootfs_ipk bbclass: special-case base-passwd preinst to run first

Preinst are run alphabetically which breaks when e.g. avahi-daemon needs /etc/passwd present.

(From OE-Core rev: d6793165feb26c51b5f19ad1e6d1a4099878e879)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Increment version post release

(Bitbake rev: 9c097704b4309dbe67fc360c8377edbedc354d00)

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

Release 1.14.0

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

codeparser: make var_expands actually hold useful information

Previously, it was calling var_expands.update() rather than add(), with
a string argument, resulting in adding each character of that string to
the var_expands set, rather than the string itself.

(Bitbake rev: 8e4e75383e43d6da2c16ec5286186a0d0569b0f8)

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

avahi: fix useradd race condition

Avahi doesn't work at boot because of:

+ sh /OE/../rootfs/var/lib/opkg/info/avahi-daemon.preinst
Running useradd commands...
grep: /OE/../rootfs/etc/passwd: No such file or directory

That is due to:

Package: avahi-daemon
Version: 0.6.30-r9.0
 [..]
Depends: libavahi-core7 (>= 0.6.30), libdaemon0 (>= 0.14), libcap2 (>= 2.22), libavahi-common3 (>= 0.6.30), libdbus-1-3 (>= 1.4.12), sysvinit-pidof, libc6 (>= 2.12), libexpat1 (>= 2.0.1)

After this patch:

 Package: avahi-daemon
 Version: 0.6.30-r10.0
 [..]
 Depends: libavahi-core7 (>= 0.6.30), libdaemon0 (>= 0.14), libcap2 (>= 2.22), libavahi-common3 (>= 0.6.30), libdbus-1-3 (>= 1.4.12), sysvinit-pidof, libc6 (>= 2.12), shadow, libexpat1 (>= 2.0.1), base-passwd

This also changes ${PN}-daemon to avahi-daemon to be consistent with the PACKAGES/FILES lines below

(From OE-Core rev: f01fbc17b5d9bf9a227d64fe858014376cd19432)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

meta: glib-2.0: don't apply qsort_r test removable patch for native version

On some buildhosts with an older version of native glib-2.0 installed (in this case
2.16.6) the qsort_r test removable patch leads to a compilation error:

| ./.libs/libglib-2.0.so: undefined reference to `qsort_r'
| collect2: ld returned 1 exit status

This patch fixes this so the patch gets only applied for the native version of this
recipe.

(From OE-Core rev: 37579d7d74d127c90c1e078d05c5bf4ba0b3f755)

Signed-off-by: Simon Busch <morphis@gravedo.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gettext: Switch to disabling nls and reducing -native build dependencies to a set of pre built m4 macros

When analysing our build performance, its apparent that binutils-cross
takes an age to get built. This is due to its dependencies on
flex-native and bison-native which in turn depend on gettext-native.

gettext-native is problematic as it has a significant dependency chain
of its own and takes an age to build. What is worse is that we never
care about the native language support in -native and -cross packages
since we always force the C locale.

This patch therefore disables nls for all -native packages (its already
disabled for -cross) and adds a new gettext-minimal-native package which
contains the m4 macros to keep autoconf/automake happy.

This means we gain a significant build time speedup by the removal of
gettext-native from most dependency chains (only being part of gettext
for the target now).

For now the LICENCE field says GPLv3, the macros are actually under a
FSF MIT like licence so we need to update this part of the patch in due
course.

(From OE-Core rev: 01b2a16beb4a924077b74943ad4d6e7976563ff1)

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

lame: add SRC_URI checksums

(From OE-Core rev: a46c6c6e8926e202aacf85a6b00dfd4e9c128792)

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

libxslt: use Copyright in LIC_FILES_CHKSUM instead of COPYING

* COPYING is replaced by symlink to Copyright during do_configure
  (see configure.in), then we end with link to nonexistent file like this:

OE om-gta02@shr ~/shr-core $ ll tmp/deploy/licenses/libxslt/
total 40
drwxr-xr-x   2 bitbake bitbake  4096 Nov  2 00:27 ./
drwxr-xr-x 818 bitbake bitbake 32768 Nov  2 00:27 ../
lrwxrwxrwx   1 bitbake bitbake     9 Nov  2 00:27 COPYING ->  Copyright
lrwxrwxrwx   1 bitbake bitbake    52 Nov  2 00:27 generic_MIT ->  /OE/shr-core/tmp/deploy/licenses/common-licenses/MIT

(From OE-Core rev: c67130e40018742463e1bd4914eb7c280425f4e8)

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

distro_tracking_fields: Update the Month to 'MMM'

Use format 'MMM' instead of spell out a Month.

(From OE-Core rev: 619408fc045e9403cd131ee823f7ac4b1230b96d)

Signed-off-by: Wenzong Fan <wenzong.fan@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dbus: ensure that the useradd shell is set to /bin/false

(From OE-Core rev: 899efe6bf17a31d842ff2f65704d4858892496d4)

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