angstrom-codec-engine-latest-preferred-versions: bump cgt to .17

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

angstrom: move to CE latest

Signed-off-by: Koen Kooi <k-kooi@ti.com>

angstrom-codec-engine-latest-preferred-versions: bump CE to 2.26.01.09

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-codecs-omapl138_4.00.00.00: add new recipe for omapl138 codec

Recipe downloads TSPA codec's from the web and packages/build them using
the configuration file from git://arago-project.org/git/projects/codec-servers.git.

Note that recipe is preparing the source after do_unpack and before do_patch -
mainly because we want to ensure that sourceipk class picks the right sources.

Signed-off-by: Brijesh Singh <bksingh@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-dvsdk-demos: sync SRC_URI with arago

Signed-off-by: Koen Kooi <k-kooi@ti.com>

qt-embedded-widgets-demo: import from arago

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-linux-driver-examples: add driver examples recipe

* This recipe will build and install the linux driver
  examples for DaVinci devices on the target.

Signed-off-by: Chase Maupin <chase.maupin@ti.com>
Signed-off-by: Brijesh Singh <bksingh@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-codecs-dm365: sync with arago

Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-codecs-dm355: sync with arago

Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-dmai: sync with arago

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-dmai_svn.bb: bump up svn rev

Signed-off-by: Brijesh Singh <bksingh@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-dmai_2.20.00.06: add updated version of dmai recipe for DM365

Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-paths: sync with arago by removing the 'packages' from LINK_INSTALL_DIR

* older dsplink recipes might need a fixup for this

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

gstreamer-ti: activate omapdmaifbsink plugin for omap3

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

gstreamer-ti: sync with arago

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-c6accel: switch to MACHINE_KERNEL_PR

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-codecs-dm365_4.00.00.00: update SRC_URI to use latest dm365 codecs

Signed-off-by: Brijesh Singh <bksingh@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

gstreamer-ti: update loadmodules.sh

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-c6accel: add missing patch

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-codecs-dm365_4.00.00.00: add new recipe for dm365 codecs

* This recipe downloads the individual codec.s from the web and puts them together to create a codec package.

* The output of codec installer is 'tar' file and we needed to extract this
  file during unpack and since  ti-eula-unpack has a logic for extracting
  tar file, but it assumes gzip version instead of regular tar. And I don't see
  any recipe currently using TARFILE variable hence updating ti-eula-unpack.inc
  to extract 'tar file'. If we come across a use case where we need to handle
  both 'tar' and '.gz' then will consider adding logic to differentiate between
  tar vs gz.

Signed-off-by: Brijesh Singh <bksingh@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-c6accel_1.01.00.01: add new c6accel version.

Signed-off-by: Brijesh Singh <bksingh@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-c6accel_1.00.00.04: add new recipe

C6Accel codec package wraps key DSP software kernels which can be invoked from
the ARM side using simple API calls. This codec can be used in a plug and play
like any other codec used for encoding and decoding audio and video streams.

Signed-off-by: Brijesh Singh <bksingh@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-dvtb: add new recipe to build DVTB (Digital Video Test Bench)

Signed-off-by: Brijesh Singh <bksingh@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-codec-engine_2.26.01.09: add new codec engine version.

Signed-off-by: Brijesh Singh <bksingh@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

ti-recipes: Switch remaining buildable ti recipes to use KERNEL_PR

* Switch recipes to use KERNEL_PR + append if they are buildable
   * All related KERNELs have higher PR than the previous local PR, so safe to change
   * Allows us to easily rebuild entire ti stack by bumping KERNEL_PR
   * Downside is, of course, that entire ti stack will get built each time we tweak kernels
      * However, most of TI stack has kernel dependence anyway

Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>
Signed-off-by: Roger Monk <r-monk@ti.com>
Signed-off-by: Koen Kooi <k-kooi@ti.com>

moko(suite,wm,wm-inf-ecore): bump SRCREV and remove AUTOREV from shr-autorev

fso-specs,libfso-glib: bump SRCREV

libfsotransport: enable palmpre-hsuart for all

* machine override without MACHINE_ARCH is bad idea
* 11:05:19 < mickey|office> palmpre hsuart support doesn't hurt
* 11:05:59 < mickey|office> palmpre support can be compiled in unconditionally for all platforms

qt4-tools-native: Add -no-freetype to the build

Avoid any dependencies on freetype, just disable it. We are
not interested in the GUI module right now, we build it to
have the tools/qdbus build, which also builds the qdbusviewer.

task-qte-toolchain-target: The QtAssistantClient is deprecated and removed in 4.7

Do not install the header files as they are deprecated and removed
from Qt 4.7 and onwards.

qt4-tools-native: Build the tools as -embedded to avoid X11 deps

We only need command line utilities and can avoid requiring X11.

uicmoc4: qt4-tools is the replacement.

qt4-tools exists for 4.5, 4.6, 4.7. There is nothing in the OE
tree that is still using the uicmoc4* target. Remove it so people
do not think that they need to use it.

meta-toolchain-neuros: Remove recipe, meta-toolchain-qte is a better replacement

meta-toolchain-qte is a good replacement for this target. Koen has
acked the removal and it opens the door to some other cleanup.

cornucopia: bump SRCREV for new alsa state files for 2.6.34

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

frameworkd: limit oeventsd workaround for buggy kernel only to 2.6.32

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

kbd: upgrade from 1.12 to 1.15.2, add native version

* native version is usefull in some kernel builds where drivers/char
  calls loadkeys (which are not available on some buildhosts)

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

Revert "wpa-supplicant : moved unused files to obsolete dir"

This reverts commit 9a5382aad4e5a3e52fa52722cf5587fdba958e9f.
This one accidentally also dragged in classes/kernel.bbclass

I will redo the commit for wpa-supplicant later.
classes/kernel.bbclass will be se a separate commit

Thanks for Paul Menzel for spotting this

Signed-off-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>

linux(-kexecboot)_git: bump SRCREV for 2.6.36-rc6

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

libx11: add 1.3.6 and 1.3.99.902, remove 1.3.5, change keysymdef_include_patch

* point to dir instead of keysymdef.h directly, because newer versions
  also check for vendor keysyms, see
  http://cgit.freedesktop.org/xorg/lib/libX11/commit/?h=libX11-1.3-branch&id=a32370c0854c8ad6f526dd997f14dbe3f466d4db

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

wifi-radar: moved files dir to more specific wifi-radar

Signed-off-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>

wireless-tools : moved unused files to obsolete dir

Signed-off-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>

wt : moved unused files to obsolete dir

Signed-off-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>

wpa-supplicant : moved unused files to obsolete dir

Signed-off-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>

unionfs : moved unused files to obsolete dir

Signed-off-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>

udev : moved unused files to obsolete dir

Signed-off-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>

gdb-cross-sdk: fix compilation on non-arm platforms

Not all platforms are arm. Not all platforms have sim/arm/Makefile for
sed. Grrr.

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Cc: Eric Bénard <eric@eukrea.com>
Cc: Tom Rini <tom_rini@mentor.com>

gpm: fix compilation due to missing struct ucred

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>

Merge branch 'master' of git://git.openembedded.org/openembedded into org.openembedded.dev

recode: use debian orig.tar.gz instead of gnu  one

recode-3.6.tar.gz was removed from gnu mirrors (moved to old-gnu).
Use recode_3.6.orig.tar.gz from Debian which is the same file.

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>

wvstreams-4.6: Remove dbus support (breaks build)

If wvstreams is configured with DBus support enabled (by not providing
--without-dbus) the the build fails at link time as libdbus cannot be
found:

| ./CXX -c dbus/wvdbusconn
| ./CXX -c dbus/wvdbusmarshal
| ./CXX -c dbus/wvdbusmsg
| ./CXX -c dbus/wvdbusserver
| linking libwvdbus.so...
| arm-angstrom-linux-gnueabi-g++: /libdbus-1.a: No such file or directory
| make: *** [libwvdbus.so] Error 1
| FATAL: oe_runmake failed
NOTE: Task failed: /SCRATCH/maniac/oe-merge-test/OE/tmp.6/work/armv5te-angstrom-linux-gnueabi/wvstreams-4.6-r0/temp/log.do_compile.6422
ERROR: TaskFailed event exception, aborting
ERROR: Build of /home/maniac/work/HYP/oe/3way_merge/recipes/wvstreams/wvstreams_4.6.bb do_compile failed
ERROR: Task 2024 (/home/maniac/work/HYP/oe/3way_merge/recipes/wvstreams/wvstreams_4.6.bb, do_compile) failed

The root cause for this behaviour is in wvstreams' configure.ac; the
weird dbus detection and library path handling breaks successive builds.

This patch, however, does not fix the configure.ac file but disables
DBus support for wvstreams at configure time. This at least fixes the
broken build.

Signed-off-by: Thilo Fromm <t.fromm@dresearch.de>
Acked-by: Paul Menzel <paulepanter@users.sourceforge.net>
Acked-by: Steffen Sledz <sledz@dresearch.de>

opie-taskbar-images: add prerm script removing alternatives

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>

libqpe-opie: copy double_name.patch from opie-mediaplayer1

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>

fortune-mod: update to 1.99.1-4 and drop host-only tools

Old debian patch (1.99.1-2) was removed from Debian mirrors. Update to
current patch (1.99.1-4). Also fix QA errors by removing host tools
from target pakages (strfile/unstr).

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>

t1lib: update to 5.1.2

t1lib 5.0.2 is so old, it was removed from main Debian mirrors. Update
to current t1lib.

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>

linux-kexecboot: move recipes in /linux

* some old version might hopefully be pruned

quilt-0.48: Use cp instead of mv for backing up patches.

* mv would not preserve symlinks but cp -a would.

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

quilt-native: Use the quilt.quiltrc from sysroot instead of /etc

This fix is only applied on build systems using linux

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

xorg: add new recipes for versions from 2010-09-28

* remove 3 old recipes for versions not pinned by any preferred-xorg-versions
* libx11 will be pushed soon

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

libxcb: add native and nativesdk BBCLASSEXTEND

* needed for libx11-native-1.3.99.902 and newer where --witout-xcb is no longer available

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

xcb-proto: add native and nativesdk BBCLASSEXTEND

* needed for libxcb-native

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

libpthread-stubs: add .inc file, INC_PR, native and nativesdk BBCLASSEXTEND

* needed for libxcb-native

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

initscripts-shr: drop ttySACx workarround and better ROOT_HOME detection

* ttySACx are fixed in kernels 2.6.32.23 and 2.6.34.7 and newer, fixed
  by kernel patch from Radek Polak
* ROOT_HOME problem found and fixed by Joachim Ott

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

qt4-tools-sdk: Add a Qt 4.7.0 version, create a .inc file for reuse.

Merge branch 'holger/qt4-4.7'

qt4: Bump INC_PR to force a rebuild with the mkspec changes

qt-4.7.0: Remove patches that cover up errors.

* These patches lack a proper commit message, e.g. no explanation
  why they need to be applied
* They remove libraries that were previously prevent
* They remove binaries that were previously installed
* The root cause appears that the -L for the qt lib directory is
  evaluated too late indicating an issue with the qmake spec.

qt-4.7.0.inc: Unpack the qte.sh to fix the qt4-embedded do_install

Fix the qt4-embedded_4.7.0 do_install by unpacking the qte.sh like
it is done in the other recipes.

qt4: Package the declarative debug packages into qt4-examples-dbg

qt-4.7.0.inc: Pick the linux.conf/g++.conf from the right directory

qt4-tools-native_4.7.0.bb: Use the mkspec files from qt-4.7.0

This needs to be streamlined that every Qt 4.x mkspec file
is found in this directory. At least this is a start into
the right direction.

qt-4.7.0: Use qmake feature to link with -rpath-link.

This is instructing qmake to use the QMAKE_LFLAGS_RPATH for the
lib directory in the build tree. It should be safe to use even
for software using the resulting qmake.

qt-4.7.0: Update the spec files for Qt 4.7

The biggest change is the QMAKE_RPATH to QMAKE_LFLAGS_RPATH
rename that broke linking for us.

qt-4.7.0: Fork the mkspec used in previous Qt versions

Some variables were renamed and we require an update.

ubifs-v2.6.24: avoid kernel error if ubifs superblock read fails

original commit comment from ubifs-2.6.git:

.get_sb is called on mounts with automatic fs detection too, so this
function should print an error if it cannot read the superblock in
debug mode only (new behaviour conforms the other fs types)

Signed-off-by: Steffen Sledz <sledz@dresearch.de>

libsdl-native: Convert to new style staging, remove do_stage

* Remove do_stage ()
* Replace tabs with 8 spaces in do_configure
* Bump INC_PR = "r3"

Signed-off-by: Noor Ahsan <noor_ahsan@mentor.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

sip-native: Convert to new style staging, remove do_stage()

* Convert do_stage to do_install
* Inherit recipe with python-dir for PYTHON_SITEPACKAGES_DIR variable
* Replace ${STAGING_BINDIR_NATIVE} with ${D}${bindir}
* Replace ${STAGING_PYDIR}/site-packages with ${D}${PYTHON_SITEPACKAGES_DIR}
* Set NATIVE_INSTALL_WORKS = "1"
* Bump PR from "ml0" to "ml1"

Signed-off-by: Noor Ahsan <noor_ahsan@mentor.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

sip-native: Run oe-stylize.py script on the recipe and modified the the recipe accordingly

* Run the oe-stylize.py script on the recipe and modified the recipe according to the output of the script.

Signed-off-by: Noor Ahsan <noor_ahsan@mentor.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

recode: Fix configure step.

Use += instead of _append to  CC/CXX otherwise configure fails

Signed-off-by: Alex Bennee <ajb@cbnl.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
xxxxxxxxxxxxx

python-epydoc_3.0.1.bb: merged native and non-native recipes

* removed the python-epydoc-native_3.0.1.bb file
* added DEPENDS_pn-python-epydoc-native = "python-native" and added DEPENDS_pn-python-epydoc = "python"
* added BBCLASSEXTEND = "native"
* added NATIVE_INSTALL_WORKS = "1"
* bunped PR

Signed-off-by: Fahad Usman <fahad_usman@mentor.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

python-pyrex_0.9.8.5.bb: merged native and non-native recipes

* removed the python-pyrex-native_0.9.8.5.bb file
* added DEPENDS_pn-python-pyrex-native = "python-native"
* added BBCLASSEXTEND = "native"
* added NATIVE_INSTALL_WORKS = "1"
* bunped PR

Signed-off-by: Fahad Usman <fahad_usman@mentor.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

oe.path: added 'recurse' argument to remove()

This makes it possible to specify whether the equivalent of 'rm -rf' or
only this of 'rm -f' is wanted.  Due to backward compatibility it
defaults to the recursive variant.

Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Acked-by: Chris Larson <chris_larson@mentor.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

gnugo_3.8: Add new recipe.

Cross-compiling patches included.

Signed-off-by: Yann Dirson <ydirson@altern.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

firefox_3.6.8.bb: Add mipsel support.

Patches taken from Gentoo and Debian.

Runs almost acceptably on my 350mhz/128mb mipsel, after a lengthy
startup delay. Should work for big endian mips without much more
work, but I can't test.

Signed-off-by: Graham Gower <graham.gower@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

machine/sh7785lcr: Add new machine for Renesas SH7785 reference board

* Machine config for Renesas SH7785 reference board.
* Add kernel config for 2.6.32 and 35.
* Add tune file for sh4a.

Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

tzcode-native: bump version to 2010m

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

tzdata: bump version to 2010m

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

SHR: drop preferred version for eglibc and gcc

* same versions are now in sane-toolchain.inc

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

sane-toolchain.inc: Make gcc 4.5 & eglibc 2.12 as default sane versions

GCC 4.5 has been tested quite a bit and so is eglibc 2.12 in SHR
and developers.

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

netbase: Add interface information for qemumips64 and qemumips64el

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Paul Menzel <paulepanter@users.sourceforge.net>

run-qemu.sh: Add new target mips64 and mips64el

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

linux_2.6.35.bb: Add config for qemumips64

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

eglibc-2.12: Configure for 64-bit mips

* Make it to install into /lib and /usr/lib instead of
  /lib64 and /usr/lib64

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

gcc-4.5: Add support to build for mips64

* Defaults to mabi=64 which means N64 is the default supported ABI
* Tweak multilib directories to emit mips64 into /lib instead of /lib64

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

binutils-cross.inc: Enable 64-bit bfd when configured for mips64/mips64el

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

conf/machine: Add qemumips64 machine configuration and tune file.

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

kernel-arch.bbclass: Add logic to account for mips64/mips64el in map_kernel_arch

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

distros: Disable gnu-style hash for mips64/mips64el

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

siteinfo.bbclass, site/: Add mips64/mips64el to siteinfo data

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

insane.bbclass: Recognise mips64 and mips64el

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

native.bbclass,package.bbclass,bitbake.conf: Make provisions for system libdir to be something else than '/lib'

* Add base_libdir_native as target libdir could be different e.g. lib32
  lib64

* Dont assume lib to be system libdir during packaging.

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