license.bbclass: cleanly handle invalid licenses in incompatible_license

This gives us an error message (bb.fatal) rather than a traceback due to the
uncaught LicenseError.

(From OE-Core rev: 7fab4cd27a2ad896218123ca82252e2b59526726)

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

license: correct re.search/fnmatch indentation

This was causing it to only obey the last of the elements in
INCOMPATIBLE_LICENSE.

(From OE-Core rev: df4a4b9792e3bd13b396b94e232c69054fdf19b5)

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

license.bbclass: check LICENSE_pkg, not LICENSE_pn-pkg

Recipes define the latter, as that's what goes into the binary packages, not
the former.

(From OE-Core rev: 6a59b294b99b05e8973b5f2d60678988e90e2de3)

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

classes/image: improve debug-tweaks ssh server configuration

Create a single postprocessing function that enables no-password logins
for both openssh and dropbear when debug-tweaks is in IMAGE_FEATURES,
changing its behaviour slightly:
* Run it regardless of whether ssh-server-* are in IMAGE_FEATURES so
  that it still takes effect if these are installed by adding
  dropbear/openssh to IMAGE_INSTALL.
* Enable it to be run from image.bbclass rather than core-image.bbclass
  so that it works for images that are using the former.

Second half of the fix for [YOCTO #2578].

(From OE-Core rev: 0937054e6e3e02565f57e60a8bdc14b0ad62e249)

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

dropbear: allow configuring blank password option at runtime

Instead of using IMAGE_FEATURES to control something within a recipe,
allow this to be set at runtime, avoiding the need to rebuild dropbear
when we want to change this option.

First half of the fix for [YOCTO #2578].

(From OE-Core rev: 313039590171456b652fa7a2f5823c9b7060b20f)

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

classes/rootfs_*: error out if postinstalls exist with read-only rootfs

If there are postinstall scripts to be run on first boot and
IMAGE_FEATURES contains "read-only-rootfs", we know this will fail on
the target device so just error out during do_rootfs. This check was
already implemented for ipk, so add it to the deb and rpm backends. Also
make all backends use bberror rather than just echo to display the
error.

Fixes [YOCTO #3407].

(From OE-Core rev: 510e377f252dece8850487c30482f83df4a89c7e)

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

iptables: upgrade to 1.4.17

patch added to fix cross-compilation issues

(From OE-Core rev: f6c7d5e0590e3e70fb435e747ffdb9fe586e7bfc)

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>

ptest.bbclass: don't add to PACKAGES if disabled

Without this, we have excessive dependencies leaking into the build. In
particular, glib-2.0 was sucking in python modules even if ptest wasn't in
DISTRO_FEATURES.

(From OE-Core rev: 7bf5ee91aa84a6fa85d6da361b64139889faa172)

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

weston: add reference Wayland compositor

This is the reference Wayland compositor.

If the "wayland" DISTRO_FEATURE is present the KMS and Wayland (nested)
compositors are enabled.

If the "x11" DISTRO_FEATURE is present the X11 compositor is enabled, and a
launcher installed.

This also ships a basic Wayland terminal.

Based on work by Alexandru Damian <alexandru.damian@intel.com>, Daniel Stone
<daniel@fooishbar.org>, and Damien Lespiau <damien.lespiau@intel.com>.

(From OE-Core rev: 94fcb80121b64974ff57b76275addaa684c559cf)

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

mesa-dri: enable the wayland-egl platform

If the "wayland" DISTRO_FEATURE is present, enable the Wayland EGL platform.
This is required by the reference Wayland compositor, Weston.

(From OE-Core rev: 0e56ecdd78212e7743e71c5d6a42ceb65c182786)

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

wayland: add Wayland protocol library

Wayland is a protocol for a client application to display user interface windows
through the use of a compositing window manager. This package include the
specific set of server-side and client-side C libraries that implement the
protocol.

(From OE-Core rev: 41538cb52f3a79c52aa37c34d8f275010b93c2c2)

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

x11-common: set XDG_RUNTIME_DIR if it isn't already set

Wayland needs this set for the communication pipes, so set a fallback in case it
hasn't already been set.

(From OE-Core rev: 263508220674321473a746a137bb5d4b21c55119)

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

bitbake: hob: Fix GTK+ and PyGtk version check error message to be more clear

Improve the error message the user sees if the versions are incorrect.

[YOCTO #3637]

(Bitbake rev: 7e3a99949358f4362876df5a82f8aeaae72c3c97)

Signed-off-by: Jessica Zhang <jessica.zhang@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hg.py: Fixed fetch failure that occurs when SRCREV is a tag.

Removed "-r REV" from hg clone invocation to fetch the entire repo,
rather than just the subset at the specified revision. This will ensure
that the specified tag exists for successive commands (e.g., the update
built on line 149.)

(Bitbake rev: f0a6261d3a8ede9ebdb6383e02cb2c2de1690640)

Signed-off-by: Jon Szymaniak <jon.szymaniak@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: fetch2: Sort file checksums by value, not path

Changing the path to a file could change the task hash even if the file
still has the same checksum. This occurs when the task depends on
multiple files and the sort order of their paths changes. Usually the
sorting is consistent because layers tend to have the same relative
paths, but this should take care of other configuations.

The problem arose when using a .bbappend to add files to a recipe in
another layer. If the layer is located alongside the other layers and
their parent directory is moved, the hash does not change. However,
moving the .bbappend layer outside of the common directory can change
the path sort order and the task hash.

(Bitbake rev: 22bd19d208f0251f5a1f9b98f3cac66181f3fc07)

Signed-off-by: Tyler Hall <tylerwhall@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake: data_smart.py and friends: Track variable history

This patch adds tracking of the history of variable assignments.
The changes are predominantly localized to data_smart.py and
parse/ast.py. cooker.py and data.py are altered to display the
recorded data, and turn tracking on for the bitbake -e case.
The data.py update_data() function warns DataSmart.finalize()
to report the caller one further back up the tree.

In general, d.setVar() does what it used to do. Optionally,
arguments describing an operation may be appended; if none
are present, the operation is implicitly ignored. If it's
not ignored, it will attempt to infer missing information
(name of variable, value assigned, file and line) by examining
the traceback. This slightly elaborate process eliminates a
category of problems in which the 'var' member of the keyword
arguments dict is set, and a positional argument corresponding
to 'var' is also set. It also makes calling much simpler for
the common cases.

The resulting output gives you a pretty good picture of what
values got set, and how they got set.

RP Modifications:

a) Split from IncludeHistory to separate VariableHistory
b) Add dedicated copy function instead of deepcopy
c) Use COW for variables dict
d) Remove 'value' loginfo value and just use 'details'
e) Desensitise code for calling order (set 'op' before/after
   infer_caller_details was error prone)
f) Fix bug where ?= "" wasn't shown correctly
g) Log more set operations as some variables mysteriously acquired
   values previously
h) Standardise infer_caller_details to be triggered from .record()
   where at all possible to reduce overhead in non-enabled cases
i) Rename variable parameter names to match inference code
j) Add VariableHistory emit() function to match IncludeHistory
k) Fix handling of appendVar, prependVar and matching flag ops
l) Use ignored=True to stop logging further events where appropriate

(Bitbake rev: f00524a3729000cbcb3317fee933ac448fae5e2d)

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

bitbake: bitbake: data_smart.py and friends: Track file inclusions for bitbake -e

This code adds inclusion history to bitbake -e output, showing
which files were included, in what order. This doesn't completely
resolve timing questions, because it doesn't show you which lines
of a file were processed before or after a given include, but it
does let you figure out what the path was by which a particular
file ended up in your build at all.

How it works: data_smart acquires a .history member, which is an
IncludeHistory; this represents the inclusion of a file and all its
inclusions, recursively. It provides methods for including files,
for finishing inclusion (done as an __exit__), and for
dumping the whole tree.

The parser is modified to run includes inside a with() to push
and pop the include filename.

RP Modifications:

a) Split Include and Variable tracking
b) Replace deepcopy usage with dedicated copy function
c) Simplify some variable and usage

(Bitbake rev: b2dda721262da8abb7dc32d019e18fbc32ed8860)

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

bitbake: BBHandler/data: Standardise some setVar access formats

(Bitbake rev: e12c1a485f96a4701144ac81179ae1af348e5bf3)

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

update-rc.d: fix failure on target

The problem happened because the variables given to -n were not
surounded by quotes.

(From OE-Core rev: 98f17da7ccc1bf79fc5894f90e52769bdbcf89df)

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

kernel.bbclass: don't depend on DATETIME

* http://cgit.openembedded.org/openembedded-core/commit/?id=67e77cd948c69798695d5d5f34163ef89339b35a
  moved DATETIME dependency from MODULE_TARBALL_BASE_NAME to MODULE_IMAGE_BASE_NAME
  but without copying that vardepsexclude

(From OE-Core rev: 9a93866067253746a889bfe444b6036cf85944b4)

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

libffi: add AArch64 support

(From OE-Core rev: de7f86532ad284f4a3c3f1486e30a3ac74763f36)

Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oprofile: add AArch64 support

(From OE-Core rev: 169b9afcf2d357fdcf254a380d21d17701685834)

Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lttng-ust: bump PE

The previous lttng-ust had a PE of "1" - that recipe was removed and
replaced with the lttng2-ust recipe, which was renamed back to
lttng-ust.  The new recipe had didn't define a PE so got the default
PE of 0 - this adds the PE back and gives it a value of "2", which
reflects the history and fixes the following error:

ERROR: Package version for package lttng-ust went backwards which
would break package feeds from (1:0.16-r1 to 0:2.1.0-r0)

(From OE-Core rev: 2fbfb2708fc7c0082b85ef964d850f8520b841d6)

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

bitbake: hig.py: refactor into individual components

Since hig.py was becoming too large (which impacted maintenance and
operations like git merges) it was split into individual files for
the classes that were implemented in hig.py. hig is now a Python
package (lib/bb/ui/crumbs/hig/). The patch was tested by
building core-image-basic/qemux86 in Hob and accessing the various
UI elements implemented in hig. Note that the patch does not change
the functionality of Hob in any way, it's just a code refactoring
that should make life a bit easier for Hob developers.

[YOCTO #3200]

(Bitbake rev: a7a2b730f915cafe7aa30539622dd1ca64ae41f5)

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

yocto-bsp: add basic git connectivity check

yocto-bsp create does a 'git ls-remote
git://git.yoctoproject.org/linux-yocto-3.4.git *heads*' to get the set
of existing branches from the kernel repo.

If the user isn't connected to the network, or if git isn't configured
sanely, yocto-bsp fails with an ugly Python backtrace.

We should try to avoid this by doing a basic sanity check for those
things before actually running the command.

The sanity check can be avoided by specifying -s on the yocto-bsp
command-line:

 $ yocto-bsp create -s test qemu

Fixes [YOCTO #3279]

(From meta-yocto rev: 496e76f9bed2ed5a04ef757724d2e63d05c7a601)

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

update-rc.d: check also that symlinks are valid

Running:

update-rc.d -r /path/to/target/rootfs basename defaults

at do_rootfs time in package postinstall stage, when
/path/to/target/rootfs/etc/init.d/basename is a symlink and points to some path
on target (for example: /etc/init.d/basename.some_package), would fail and the
postinstall execution would be postponed for first boot, on target.

This patch adds the posibility to verify whether the file the symlink
points to actually exists in the target rootfs.

[YOCTO #3716]

(From OE-Core rev: 4b63e73422ea25aba1bde0beddb02bc04948e13c)

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

apmd: fix license segment md5sum boundary

Replaced incorrect "firstline" with correct "beginline" for telling
where license segment in file begins. Old md5sum was calculated from
the beginning of the file, not from beginning of the license segment.

(From OE-Core rev: 0ada16fe509d9a8608c05d16666243c8ebe4d3b3)

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>

dbus: fix license segment md5sum boundary

Replaced incorrect "firstline" with correct "beginline" for telling
where license segment in file begins. Old md5sum was calculated from
the beginning of the file, not from beginning of the license segment.

(From OE-Core rev: 1eb8f569fa084f7f5d284f0bb1110721b9674d9a)

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>

libidn: fix license segment md5sum boundary

Replaced incorrect "firstline" with correct "beginline" for telling
where license segment in file begins. Old md5sum was calculated from
the beginning of the file, not from beginning of the license segment.

(From OE-Core rev: 3a116ba892f63a690932d0aee454c15927ba0eca)

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>

gperf: fix license segment md5sum boundary

Replaced incorrect "firstline" with correct "beginline" for telling
where license segment in file begins. Old md5sum was calculated from
the beginning of the file, not from beginning of the license segment.

(From OE-Core rev: 9851d4bdf9283d280d8fc1302f5b744b6d792759)

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>

ethtool: fix license segment md5sum boundary

Replaced incorrect "firstline" with correct "beginline" for telling
where license segment in file begins. Old md5sum was calculated from
the beginning of the file, not from beginning of the license segment.

(From OE-Core rev: da826787f08adec6a3ec3995e7809b75df2c99c3)

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>

shadow: fix license segment md5sum boundary

Replaced incorrect "firstline" with correct "beginline" for telling
where license segment in file begins. Old md5sum was calculated from
the beginning of the file, not from beginning of the license segment.

(From OE-Core rev: 9aefa35859413a635fb4e122d59a703e95a9b5bb)

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>

iptables: fix license segment md5sum boundary

Replaced incorrect "firstline" with correct "beginline" for telling
where license segment in file begins. Old md5sum was calculated from
the beginning of the file, not from beginning of the license segment.

(From OE-Core rev: bc36eadd5dee7390977629358cab3f13d6ddcf5c)

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>

gzip: fix license segment md5sum boundary

Replaced incorrect "firstline" with correct "beginline" for telling
where license segment in file begins. Old md5sum was calculated from
the beginning of the file, not from beginning of the license segment.

(From OE-Core rev: 029965a3e8263ec11d9a5f1bfbb8152ae4a1aea3)

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>

libidn: add checksums to version 0.6.14 recipe

(From OE-Core rev: 98691df00020036628617d6f149c02b1d0d856ea)

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>

ofono: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 1766f851be890ec057040594a12997b1576ba1d5)

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>

tremor: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: e64800e605635737b4d74a36e66b0f7e0fee08bc)

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>

alsa-utils: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 723ec3b3b618fc4cccbd0e7306624161dd81b0e5)

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>

bluez-hcidump: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: cb9480937162c4e6a80769f5ccd7cbc09d8947d2)

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: fix build with automake-1.13

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.
Remove AC_CONFIG_MACRO_DIR override that caused aclocal to fail
since it didn't get m4-directory information.

(From OE-Core rev: 54a945a5e67c44be3e0b21114b7ca3abf8d99ada)

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>

libtirpc: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 074566f49be59b2f507ba427f258e1ca345da3c9)

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>

vte: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 4cd77250aa65b18c863a112595bc30391b23d032)

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>

sysfsutils: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 11319ad6eec2056278d6b6540e31d369870537da)

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>

rgb: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: b1c788dbd7b0f015718f915763b1d08e566e80e3)

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>

libxsettings-client: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 703ea77f1869c1101e1637831a83ca2d754907da)

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>

pango: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 16df16408719b93f595412a4c10df793d847dd19)

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>

polkit: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 9f4cca12f00327bb9a7070f305c36433ee732d7a)

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>

image_types.bbclass: Add and update comments regarding image types.

Explain the elf/cpio.gz snippet of code, and update a comment to
reflect the correct variable name.  No functional change.

(From OE-Core rev: 0ca4a3a25e4495bf6e2e3932c019bd15856faf55)

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

connman: upgrade to 1.10

0002-storage.c-If-there-is-no-d_type-support-use-fstatat.patch
    - adapted to the new version

0001-timezone.c-If-there-is-no-d_type-support-use-fstatat.patch
    - patch removed (it is included in the new version)

inet-fix-ip-cleanup-functions.patch: added
    - fix for ip cleanup functions

(From OE-Core rev: e0cca947fc87c8f7a6d2a8fd719c87d56dd52941)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
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>

connman-conf: configures connman in qemu machines

For qemu machines, connman should not configure
virtual wired interfaces. Wired interfaces (eth0)
are assigned static IP addresses by the virtual machine
manager (qemu). This packages places the eth0 interface in
a list of blacklisted interfaces, forbiding connman
administering wired interfaces for qemu machines.

(From OE-Core rev: cfd9ab0eff27a59b8b9e750a01565e2e7c94e801)

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>

Removed undisplayed picture from connman-gnome

- removed the picture that doesn't display in the
connman-gnome "about" section.
- updated the recipe to use the patch when running bitbake
- created a folder with patches named "connman-gnome",
the same as for connman.

(From OE-Core rev: 56c7ce56a1c9870df08821c6f18fac3713b1b4cf)

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>

runqemu scripts: add support for booting an ISO image

Add support for booting an ISO image for runqemu scripts.

[YOCTO #3710]

(From OE-Core rev: edd629029979dc18905fce23b64216f15eea501f)

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

recipes-kernel/lttng-2.0: rename to recipes-kernel/lttng

recipes-kernel/lttng was removed by a previous patch, which also
removed the naming conflict that caused the lttng-2.0 version of lttng
to be named recipes-kernel/lttng-2.0.  Since the naming conflict has
disappeared, we can go back to the simpler naming.

(From OE-Core rev: 2027fd1bbfc2c136672d5054ba722e9d9d80451d)

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

lttng2-ust: rename to lttng-ust

recipes-kernel/lttng/lttng-ust was removed by a previous patch, which
also removed the naming conflict that caused the lttng-2.0 version of
lttng-ust to be named lttng2-ust.  lttng2-ust was the only lttng-2.0
recipe to be named in this way and since that was entirely due to the
conflict, we can go back to the consistent naming for it.

(From OE-Core rev: 3c927c39502061bdf3ef9fba1f0f6e7080f6c3cd)

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

lttng: remove 'legacy' lttng

This removes 'legacy' lttng - it hasn't actually worked for awhile
now, and has been replaced by equivalent functionality in
recipes-kernel/lttng-2.0 (which will be renamed to lttng in a
follow-up patch).

Fixes [YOCTO #3303].

(From OE-Core rev: 920b4c292f6c704cfe32666a3e1de5bce19043b5)

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

quota: upgrade to 4.01

(From OE-Core rev: 879c36e69889ed4945f427d5947555949c024023)

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>

xz: LICENSE field is wrong

xz contains a GPLv3 m4 macro which is used at build time, this does not make
the installed or runtime items GPLv3, which is what the LICESNE field pertains
to, therefor remove it from the LICENSE field.

(From OE-Core rev: d1950aaed812dc98d0536ec819c82b7a6160b316)

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

iproute2: upgrade to v3.7.0

configure-cross patch is now simplified.
sysroot is now properly set.

(From OE-Core rev: 9c9c55635ef34fc0f4d5acd8434c8f64ccd5bf87)

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>

distrodata: Updated checkpkg task

bitbake -c checkpkg fetches the latest version for all packages,
including the ones from sourceforge and those that have
ununsual package naming.
Also removed all pn-native/nativesdk-pn entries from the resulting
checkpkg.csv file if the system already reported the latest version
for pn.

(From OE-Core rev: dc33d69df55ad8d63024d34e35a0b987e01b0077)

Signed-off-by: Emilia Ciobanu <emilia.maria.silvia.ciobanu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

db: upgrade to 5.3.21

License file changed the year,
triggering a change in checksum

(From OE-Core rev: c1fe80b5b6e3ac3da8a0a1c138649c92ca11f285)

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>

add meta-toolchain-qt

- this allows the creation of a SDK containing all the needed
libraries to develop QT4 for X applications
- the recipes were factorized to share as much code as possible
- binaries generated using the toolchain were run tested on an
ARMv7 target
- both qt & qte toolchains were build tested for qemuarm target
and for an ARMv7 target

(From OE-Core rev: cc803b06841e75aa6591abaeda0a788e888e208b)

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

qemu: machine override ability added

Add override ability to qemu to allow qemu specific
configurations for any qemu machine.

(From OE-Core rev: 4ee668a558e5d4a6d14e29c9fe88b8bb642a16a9)

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

ia32-base.inc: Add more macros for xf86 drivers

This commit adds macros for fbdev & modesetting X driver packages.

(From OE-Core rev: c412a4eaf17fa29a4f2280f8be1dcf95346217a6)

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

xf86-video-modesetting: new recipe for gma500

I am trying to use gma500 driver for the noemgd BSPs. The current best
xf86 driver for it is modesetting. This recipe provides that
driver.

(From OE-Core rev: 5f9de6f97cf218973a5284feb1e5ae841b69dcc6)

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

perl: fix installed but not shipped issue

It shows warning when bitbake perl:
WARNING: QA Issue: perl: Files/directories were installed but not
shipped
  /usr/lib/perl/5.14.2/auto/XS/Typemap

That because file Typemap.so is not install correctly. Fix it.

(From OE-Core rev: d4c3cc8cff9311ff873304410e3092921dbe6ddc)

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

connman: fixed init script so connman can runs over nfs

Adding -I to the arguments, connmand will ignore the eth interface, so if you
are using nfs it will not be disconnected.

OBS: it might not work if using more than one eth interface. But it's better
than not using connman with nfs at all.

(From OE-Core rev: 6080d1dc719c38fd97c2abd38e0ad938fbf6f452)

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

initscripts: remove finish.sh

Remove finish.sh from initscripts as it is no longer used.

(From OE-Core rev: 3bcd9188f919da1b10f692d29ef34cd18ef88184)

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

gitignore: only ignore meta- directories

else for example meta-toolchain-qte.bb is also ignored.

(From OE-Core rev: d15d42562dff08696015f4d15e3e15c70ee2357a)

Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>

quilt: Don't use BUILD_ROOT on darwin.

(From OE-Core rev: 130f1fa3d2787c3c542522ecf9ba5debf914f41a)

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

quilt: Remove non-gnu.patch, and added configure flags for darwin.

(From OE-Core rev: f495dfb60cb1fd86b7365210dd7aab75cec2d5e4)

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

sstate: Do not add the --no-run-if-empty arguement to xargs when on Darwin, as it is not supported.

(From OE-Core rev: 73b34d43633717b99e4f2f669939939cfa59eecb)

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

gcc: add missing dependency (zlib)

- add zlib to DEPENDS (it fails otherwise)

(From OE-Core rev: 7a0fd7145461d3feb9fe7900a39376d556e655d6)

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

bluez4: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: e4e3fc72fdfc7cde4652c562a936ef473c35367a)

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

gmp: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 17a58db25657184e0770d29391062e7f93d5bb5b)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Conflicts:
meta/recipes-support/gmp/gmp_5.1.0.bb
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gtk+: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 0beee6281fed3926b6871dbfdd907416d4a04886)

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>

startup-notification: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: f311b2d9978e07d4e5f4b6687ff2dba41b5e5da4)

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>

gconf: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 388939c4cebabce8ef05c8aec5b9dccf390b75f0)

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>

opensp: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: ae3917b9dc8c350de92b9ec0c7def780f8a09af7)

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>

libvorbis: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 0f2a601b04ac564905942b4d2006dbf348b1517b)

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>

libmad: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: da623c24804b7bb331de45f39063c7b207d6e4d6)

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>

libogg: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 8bfb260eb601360bea251d77633f463063e4d451)

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>

python: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch for both python-dbus and
python-pygobject that replace automake macros no longer supported
by automake-1.13 with modern constructs.

(From OE-Core rev: 9866bed92c0f3dee3f1674e9f9a28e21f5ba33fc)

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>

alsa-lib: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 453fd36914fe1ae41a2115266833ec7fcba0869d)

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>

makedepend: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 6cb2ef796278b0747cda4771091d9fdc1603bea7)

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>

tslib: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 4530d308a546e989873e2978ea5ed2970e8925e7)

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>

dbus-glib: replace obsolete automake macros with working ones

Add obsolete_automake_macros.patch that replaces automake macros
no longer supported by automake-1.13 with modern constructs.

(From OE-Core rev: 3f5e1d92a53015b44fa4eaf5325ba742a9a854e6)

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>

help2man: update to upstream version 1.40.13

(From OE-Core rev: dd1e33420cee433dc5c4bd302e44ccfc37eb6722)

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>

coreutils: fix license segment md5sum boundary

Replace incorrect "startline" with correct "beginline" for telling
where license segment in file begins. Old md5sum was calculated
from the beginning of the file, not beginning of the license segment.

(From OE-Core rev: 98c258f5db7e6800557e4e51c064b0d82a10edd6)

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>

libxdamage: update to upstream version 1.1.4

Compatible with automake-1.13.

(From OE-Core rev: c50d7a2869f4d9004bc39444d798a075607d03fa)

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>

linux-libc-headers: fix headers install in long path name environments

If a build is started in a deep directory structure, or in a path with
long directory names the installation of linux-libc-headers will fail
with:

  | make[2]: execvp: /bin/sh: Argument list too long
  | make[2]: ***

>From within the kernel Makefiles themselves.

The solution is to patch the kernel build rules with the following
change:

[
    scripts/Makefile.headersinst: install headers from scratch file

    If headers_install is executed from a deep/long directory structure, the
    shell's maximum argument length can be execeeded, which breaks the operation
    with:

    | make[2]: execvp: /bin/sh: Argument list too long
    | make[2]: ***

    By dumping the input files to a scratch file and using xargs to read the
    input list from the scratch file, we can avoid blowing out the maximum
    argument size and install headers in a long path name environment.
]

Until this change, or a similar one, is merged into the mainline kernel, this
change should be applied to any 3.7 or greater libc-headers build.

Upstream-status: Pending

(From OE-Core rev: 1e63a3b7b7915d40bb59976a02b9f53968997ed3)

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

gtk+3: add 3.4.4

Mostly taken from meta-gnome, with some changes.

Instead of depending on gtk+-native, depend on just gdk-pixbuf-native and build
the tool required for the host directly.  This saves building the entire GTK+
stack natively just for one tool.

Append a version suffix and use update-alternatives on gtk-update-icon-cache as
with GTK+ 2.

(From OE-Core rev: 16c0556808029a0f6a3f91a726b14626e9489498)

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

gtk+: version installed gtk-update-icon-cache and use alternatives

With GTK+ 3 being added soon we'll have potentially two providers of
gtk-update-icon-cache.  Append a version to the binary and use
update-alternatives to ensure that the unversioned name works.

For gtk+-native the alternatives won't take effect, so install a symlink
explicitly.  This will break if we have a gtk+3-native but we can fix that if
and when that happens.

(From OE-Core rev: a9e71b41463c6c8e2be50d40403a0017d3cdf146)

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

boost: Use PARALLEL_MAKE for bjam

Like make, bjam accepts the parameter -jX for multithreaded execution
do_install also profits from this setting
Tested with a quad core 64bit intel cpu (with hyperthreading) PARALLEL_MAKE="j16"

$ time bitbake boost

before:
real    14m37.433s
user    7m40.785s
sys     4m30.109s

after:
real    7m11.979s
user    12m10.694s
sys     2m47.078s

Also fixes tab indention

(From OE-Core rev: 9b9cfc1dfe5e3b8f89b7a8508537166d0f23935e)

Signed-off-by: Samuel Stirtzel <s.stirtzel@googlemail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

puzzles: upgrade to r9733

(From OE-Core rev: 418312a2523f5e1fd231d215475250168f17fa58)

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

alsa-utils: add bash as a dependency of the alsa-utils-alsaconf rpm.

alsa-utils-alsaconf rpm contains the bash script alsaconf. In order
to have the do_rootfs not error when bash is not brought along in
the rootfs as an explicit dependency, we add it as a dependency
of alsa-utils.

(From OE-Core rev: 18575b082a4042376fd1575465e69562dea04ddc)

Signed-off-by: Florin Sarbu <florin.sarbu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

apt-native: fix the creation of apt.conf.sample

1.The file of "apt.conf.sample" is in the outdir, and outdir is assigned
by "os.path.join" with the params of ${D}, ${sysconfdir} and "apt". But
${sysconfdir} is an absolute dir and that let ${D} be discarded.

The following is the help on function os.path.join(a, *p):
     Join two or more pathname components, inserting '/' as needed.
     If any component is an absolute path, all previous path components
     will be discarded.
Use oe.path.join instead which don't have this problem.

2. Create apt.conf.sample if it doesn't exist.

[YOCTO #3677]

(From OE-Core rev: 95a655a853b7cd43163362d72da0c134d4c0ec85)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pango: update to upstream version 1.32.5

multilib-fix-clean.patch adjusted to apply cleanly, and to handle
new modules.cache file.

(From OE-Core rev: bcc5773199b46811f913d06e960d34ed9833f714)

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: add recipe, version 0.9.10

OpenType text shaping engine HarfBuzz is dependency of pango >= 1.32
This depends on all of glib-2.0, cairo, and freetype to consistently
provide all backends. Otherwise it would, in typical case, depend
on build order of these packages. HarfBuzz configure would pick those
backends for which dependencies happen to be built before it.
Similarly graphite is consistenly disabled if there ever happens
to be recipe for it.

(From OE-Core rev: 6f5ab158203101db69d11754a77b0fd6fbb23f87)

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>

directfb: fix build against Mesa 9

Mesa 9 changed the libgbm API, so take a patch from upstream to fix this.

(From OE-Core rev: ab6f8ea9ddc69433c3e84058d8ea5502cbab35c3)

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>

mesa: bump to 9.0.1

Upgrade Mesa to 9.01, backporting a patch from master to build without needing
mesa-dri-glsl-native.

Several patches have been merged upstream, so drop them.

License checksums updated as the file was re-formated.

Drop mesa-dri-glsl-native as we don't need it anymore.

Tested with Piglit's "quick" test suite, pass rate remains constant compared to
8.0.5 at 87%.

(From OE-Core rev: f816697c1e76589de7853bd5573cf578e25c7ca7)

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>