xorg-app: obsolete old versions

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

xorg-lib: obsolete old versions

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

xorg-xserver: obsolete old versions

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

xorg-util: obsolete old versions

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

xorg-driver: obsolete old versions

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

xorg-proto: obsolete old versions

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

xorg-doc: obsolete old versions

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

xorg-font: obsolete old versions

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

xorg-data: obsolete old versions

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

packaged-staging.bbclass: bug fix in ipk file name generation for searching ipk file.

* First check the availability of PKGV_pkg variable, if it is not available then get value from PKGV variable for creating ipk file name for searching ipk file.

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

linux-kexecboot: bump PR to follow kexecboot upgrades.

Signed-off-by: Andrea Adami <andrea.adami@gmail.com>

kexecboot: bump to e951750b09eb0e1ba81f1bf71d6108c8d3288f91

* sync with upstream
* new configure option to add a delay before device detection
* Zaurus spitz needs 2 secs because of slow internal HD init (with empty mtd's)
* bump PR

Signed-off-by: Andrea Adami <andrea.adami@gmail.com>

eglibc-options.inc: Account of DISTRO_FEATURES_EGLIBC being not defined.

* Thanks Bernhard

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

ti-linux-driver-examples_git.bb: Remove deprecated usage of 'patch' url param

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

devmem2: fix patch

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

argp-standalone: build shared library also

This applies a patch from FreeBSD to use libtool to build both shared and
static versions of libargp.

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

eglibc: Configure based on DISTRO_FEATURES_EGLIBC

* Let distros decide what features to enable if not
  selected it will turn off all possible knobs

* Enable all knobs for angstrom and distros based on
  sane-toolchain

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Martin Jansa <Martin.Jansa@gmail.com>
Acked-by: Koen Kooi <koen@openembedded.org>
CC: koen@dominion.thruhere.net
CC: koansoftware@gmail.com

gcc-4.5: Get few more fixes from linaro and upstream gcc-4_5-branch

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

linux-omap 2.6.37: check in updated overo patch

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

ti-cgt: add 7.0.4

This adds support for mini64 aka 'Tesla'

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

linux-omap: move to 2.6.37 final

There is no 2.6.37-omap1 tag and linus' tree is missing fixes, so patch up -rc8 with the patches from mainline

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

u-boot-sakoman: move to 2010.12

Tested on pandaboard EA1 and A1 ES2.0 boards

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

cornucopia.inc: bump SRCREV for latest fixes

Signed-off-by: Klaus Kurzmann <mok@fluxnetz.de>

devmem2: remove comment of word write from fsl patch

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

devmem2: add freescale alignment fixup patch

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

rng-tools: dep on argp-standalone for uClibc

argp is hierarchical argument parsing built on top of getopt, but is part of
glibc.  Niels Möller created a standalone version, so make rng-tools depend
upon that when we're using uClibc.

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

e-wm: bump SRCREV a bit more than EFL_SRCREV is now

* ie illume-keybard makes E segfault in some configurations

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

apache2: Fix invoking buildconf

It may not have execute set so start it via sh

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

site/powerpc-common: Drop ac_cv_func_dlopen

This is something that can be done at compile time and more importantly
was set wrong (this should be no and then another one yet for
getting -ldl, but that also wouldn't live in the -common one).

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

boost-1.45.0: force ARM mode for now, thumb makes ld to buffer overflow

* can be removed as soon as binutils are fixed
* testcase is here http://build.shr-project.org/tests/jama/binutils.overflow.test.tgz with failing ld call included in ld.call.sh

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

elmdentica: add patch for azy api change from r56034

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

mesa: add new 7.10 version, with D_P -2

* someone please test uclibc build if old uclibc.patch is enough

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

ntp-ssl: fix PACKAGES

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

e-base: bump EFL SRCREV a bit again to drop libeflvala patch

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

vdr: update to 1.7.16 from 1.7.10

Several of Hennig’s patches were applied upstream and are not necessary anymore.

• `recipes/vdr/files/cplusplus.patch`: replaced by `recipes/vdr/vdr-1.7.16/0002-tools.h-include-stdarg.h.patch`
  1. `canonicalize_file_name()` is available in uClibc now.
  2. Deprecated C headers will remain in VDR instead of C++ headers [1].
• `recipes/vdr/files/disable_plugin.patch`: removed, since ncurses is found correctly now.
• `recipes/vdr/files/fixpaths.patch`: removed, since `INCLUDES` is not overwritten anymore since VDR 1.7.15 [2] and options can be passed from outside since 1.7.13 [3].
• `recipes/vdr/files/libintl.patch` ? `recipes/vdr/vdr-1.7.16/0003-uclibc-Add-lintl-to-LIBS.patch`
  1. Still needed. I do not know how to fix this upstream.
• `recipes/vdr/vdr-1.7.10/linkerflags.patch` ? `recipes/vdr/vdr-1.7.16/0001-Makefile-Pass-LDFLAGS.patch`
  1. Still needed and extended to cover all places to get rid of `ERROR: QA Issue with libvdr-svcsvr: No GNU_HASH in the elf binary:`. This patch will not be needed anymore in 1.7.17 [4].

Since 1.7.x are development releases and no users of VDR in OE are known to me simply replace 1.7.10 by 1.7.16.

This recipes is build tested for `MACHINE = "beagleboard"` with `angstrom-2008.1`, `angstrom-2010.x`, `minimal`, `minimal-uclibc`.

[1] http://www.linuxtv.org/pipermail/vdr/2010-December/024034.html
[2] http://www.linuxtv.org/pipermail/vdr/2010-April/022831.html
[3] http://www.linuxtv.org/pipermail/vdr/2010-February/022333.html
[4] http://www.linuxtv.org/pipermail/vdr/2010-December/024096.html

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

Convert `libcap_1.10` to new staging system by removing `do_stage()` [1].

This change is build tested with the following configuration and `PREFERRED_BINUTILS_VERSION = "2.21"`.

        Build Configuration:
        BB_VERSION        = "1.10.2"
        METADATA_BRANCH   = "org.openembedded.dev"
        METADATA_REVISION = "224c193"
        TARGET_ARCH       = "arm"
        TARGET_OS         = "linux-uclibceabi"
        MACHINE           = "beagleboard"
        DISTRO            = "minimal-uclibc"
        DISTRO_VERSION    = "dev-snapshot-20110109"
        TARGET_FPU        = "soft"

[1] http://wiki.openembedded.org/index.php/Legacy_staging

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

libcap, libcap2: use automatic packaging to fix QA errors

Since commit 6883f6 [1] packages have been populated manually.

FILES_${PN} = "${libdir}"
FILES_${PN}-dev = "${includedir}"

This gives the following QA errors for `libcap` and `libcap2` (which is based on `libcap` [2]) [3].

• `libcap`:

ERROR: QA Issue with libcap: non -dev package contains symlink .so: libcap path '/work/armv7a-oe-linux-uclibceabi/libcap-1.10-r2/packages-split/libcap/usr/lib/libcap.so'
ERROR: QA Issue with libcap: non -dev package contains symlink .so: libcap path '/work/armv7a-oe-linux-uclibceabi/libcap-1.10-r2/packages-split/libcap/usr/lib/libcap.so'

• `libcap2`:

ERROR: QA Issue with libcap2: non -dev package contains symlink .so: libcap2 path '/work/armv7a-oe-linux-uclibceabi/libcap2-2.16-r1/packages-split/libcap2/usr/lib/libcap.so'
ERROR: QA Issue with libcap2: non -dev package contains symlink .so: libcap2 path '/work/armv7a-oe-linux-uclibceabi/libcap2-2.16-r1/packages-split/libcap2/usr/lib/libcap.so'

Using automatic packages the symlink `libcap.so -> libcap.so.1.10` is put into `libcap{,2}-dev`.

Note: The symlink `libcap.so.1 -> libcap.so.1.10` is still put into `libcap1_1.10-r2_armv7a.ipk` but does not raise a QA error.

[1] http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=6883f65588c6aa7e9780ff7c11b15a397382a422
[2] http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=c71a6ebf796df907ba0d4339e6acf66faa3c6232
[3] http://lists.linuxtogo.org/pipermail/openembedded-devel/2011-January/028557.html

Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
CC: Rolf Leggewie <no2spam@nospam.arcornews.de>
Acked-by: Dirk Opfer <dirk@do13.de>
Acked-by: Dirk Opfer <dirk@do13.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

xz:  Configure with --enable-small --disable-lzma-links

Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>

fso-autorev: drop libeflvala as it's in EFL repository now

* and SRCREV there is changed too often without any libeflvala change

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

slugos.inc: Use bluez4 as default provider for bluez-libs

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

linux-ixp4xx.inc: Add DEPENDS on zd1211-firmware

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

gcc-svn: remove unrelated hunk from patch

It provokes conflicts later on, so drop it since it serves no real use.

Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>

patch.bbclass: conditional patch-tool dependency

only depend do_patch on patch-tool if recipe references any patch

The whole of do_patch should be gated through a facility to determine if
there is any patch at all. TODO.

Acked-by: Leon Woestenberg <leon@sidebranch.com>
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>

gcc: replace unconditional SRC_URI_append with SRC_URI +=

* Fixes missing leading whitespace character at some locations

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
CC: Khem Raj <raj.khem@gmail.com>
Acked-by: Khem Raj <raj.khem@gmail.com>

ltrace: Rework sysdeps dependencies

The dependencies here were slightly lazy and in certain cases it was
possible to be building objects that required generated headers before
they were done.

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

autoconf: Really only build in desired SUBDIRS

In certain cases we would re-run configure as part of do_compile
and thus wipe out our changes.  So lets make the change before
we run configure the first time so they are preserved.

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

alsa-lib: fix packaging error

* symlink libasound.so -> libasound.so.2.0.0 in /usr/lib was missing in
  dev package

Signed-off-by: Steffen Sledz <sledz@dresearch.de>
Acked-by: Koen Kooi <koen@openembedded.org>

scons: add EXTRA_OESCONS

It's similar to EXTRA_OEMAKE, which appends arguments to the scons
command

v2:
* initialise the EXTRA_OECONS if it is not defined

Signed-off-by: Víctor Manuel Jáquez Leal <vjaquez@igalia.com>
Signed-off-by: Koen Kooi <koen@openembedded.org>

fsodeviced: i2c-tools no longer a dependency; everything necessary included in fsodeviced itself

linux-2.6.34: upgrade to longterm 2.6.34.8 and update openmoko and shr patch a bit

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

ntp: cleanups

* remove obsolete versions
* move ntp-ssl forward
* fix ifup script
* fix crontab
* moved files into files/ so it can be shared between ntp and ntp-ssl

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

linux-2.6.34: remove c7x0 patch from SRC_URI override as that file was removed in f000274e8bcae7c2d8cfebb2a7485689e7cab476

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

busybox: remove 1.17.4

now that angstrom-2010 is using 1.18.1

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

angstrom-2010: prefer busybox-1.18.1

this is the latest stable version

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

tslib-git: update to git master

* 5243db505b109df3001f doesn't exist so fetch is failing.
error: pathspec '5243db505b109df3001f' did not match any file(s) known to git.

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

busybox-1.18.1: add latest fixes

* tftp: fix bad interaction betweel poll() and alarm(). Closes bug 3061
This was breaking timeout handling.
http://git.busybox.net/busybox/commit/?id=84dba9c5bbd99cb80c0e201bbffa27a51766c63f

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

linux-libc-headers_2.6.37: fix SRC_URI

actual SRC_URI can't work because :
- linux-libc-headers-2.6.37.tar.bz2 or linux-libc-headers-native-2.6.37.tar.bz2
don't exist on kernel.org's FTP
- there is no file://defconfig

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

mesa-7.9.1: Fix build on uclibc

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

cramfs: add cvs version

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

procinfo-ng: new recipe, version 2.0.304

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

bitbake.conf: fix quote usage in COMBINED_FEATURES

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

slugos.conf: Use thumb mode to save some more space

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

libnl2: Fix a race on route/pktloc_syntax.h

At issue is that route/pktloc.c (not generated) depends on
route/pktloc_syntax.h (generated).

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

qemu.inc: Switch to xserver-xorg

Keeping with xserver-kdrive means that we have a sometimes race
in that xserver-kdrive is too old for the other xorg components that
are built (with X11R7.5) to build.  So we switch to xserver-xorg and
appropriate modules.  Tested on qemux86 and x11-image.

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

minimal.conf: Fully switch to bluez4

We still had bluez-libs being set for bluez-libs rather than bluez4 which
meant that both were being built.

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

xorg: upgrade to X11R7.6 and remove preferred-xorg-versions*.inc files

* X11R7.6-RC1 was already released
  http://www.x.org/releases/X11R7.6-RC1/doc/xorg-docs/ReleaseNotes.html
  so it's IMHO time to start migration there
* preferred-xorg-versions-live.inc was tracking latest stable released
  tarballs, but only reason I was maintaining it was to overwrite older
  preferred-xorg-versions-X11R7.5.inc if your distribution included ie
  X11R7.5.inc and you wanted live.inc from local.conf
* default recipe versions without any preferred-xorg-versions include
  are now the same as live.inc
* we will get better test coverage if we use same versions
* if someone really needs older release then pinning only core stuff
  like xserver/libx11 will be probably better, than pinning every xorg
  app/proto/font etc.
* I can slow down adding newest xorg stuff or ie keep it everytime for
  at least a week on ML for another review and test on more targets
  (I'm testing it only on om-gta02/spitz/nokia900)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Acked-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
Ackey-by: Tom Rini <tom_rini@mentor.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

task-shr-feed: add emtooth2

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

emtooth2: add recipe for emtooth2 (new vala version of the app)

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

mokophone: SRCREV bump

Reject button now acts also as mute button if ringtone is playing

Signed-off-by: Daniele Ricci <daniele.athome@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

libphone-ui-shr: bump SRCREV

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

mesa-demos: import recipe from poky

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

libglew: import 1.5.7 from poky

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

mesa: drop 7.8.2 version

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

angstrom-2010: prefer newer mesa 7.9.1

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Acked-by: Koen Kooi <koen@openembedded.org>

mesa-7.9.1: change configure.ac to check for python not python2 as python-native does not stage python2

* should we fix it in python recipes instead? maybe more projects are looking for python2 and finds is on host not sysroot

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

mesa: upgrade 7.9 to bugfix release 7.9.1

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

libeflvala: adapt to elementary API change from r55869

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

iliwi: add patch to build with newer EFL

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

e-base: bump EFL_SRCREV

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

linux-2.6.37: update openmoko and shr patch, add support for om-gta01

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

SHR: drop PREFERRED_PROVIDER_virtual/libusb0 and use default libusb-compat

* libusb has legacy staging
* libusb-compat is better maintained and newer

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

gst-tracelib: add git version

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

busybox-1.18.1: add latest fix

modprobe-small: support compressed modules in insmodmaster
http://git.busybox.net/busybox/commit/?id=e9d12b57bfb941e754d09a7d5f54c183ca7dbb25

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

fso-specs,lifso-glib: bump SRCREV

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

linux-2.6.37: add om-gta02 support

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

azy: add mysql5 to depends it's used in azy/src/tests/T_SQL.azy_server_stubs.h if mysql-config is found

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

cornucopia.inc: bump SRCREV to get fixed fsodevice.conf for gta02

Signed-off-by: Klaus Kurzmann <mok@fluxnetz.de>

preferred-slugos-versions.inc: Lock newer versions of linux-libc-headers and dbus

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

linux-kexecboot: bump PR to follow kexecboot changes.

Signed-off-by: Andrea Adami <andrea.adami@gmail.com>

kexecboot: bump to 92403e305a57154c7c779d6d47600387116a7988

* update list of configure args
* bump PR

Signed-off-by: Andrea Adami <andrea.adami@gmail.com>

qemu_0.12.5.bb: Backport patches from upstream to fix qemu for SH4

* Now we can boot SH4 qemu again with these patches and gcc 4.5 compiled
  kernel/root file system

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

uclibc_git.bb: Add patches for sub arch variant removal.

* Also Bump to latest revision

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

apex-env_1.5.14.bb: Fix compile error discovered with gcc 4.5

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

xbmc: add `python` to `DEPENDS`

XBMC needs Python to be built properly [1][2]. Python is pulled in implicitly [3], so no error is seen. Make it explicit though to be safe.

[1] https://github.com/xbmc/xbmc/blob/e9e7990b428863b8f2daad4dc5fd2b0a320aa965/Makefile.in#L475
[2] http://cgit.openembedded.org/cgit.cgi/openembedded/tree/recipes/xbmc/xbmc/0006-Hardcode-python2.6-for-now.patch?id=0cc02ded7bd429208ccb0e0e0aa3e43aec8a4722
[3] http://cgit.openembedded.org/cgit.cgi/openembedded/tree/recipes/boost/boost-with-bjam.inc?id=d470cfde5bd1d4c9ab23a37d286456726711229f#n67

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

cornucopia.inc: bump SRCREV

Signed-off-by: Klaus Kurzmann <mok@fluxnetz.de>

libfso-glib: bump SRCREV

Signed-off-by: Klaus Kurzmann <mok@fluxnetz.de>

sane-toolchain.inc: GCC 4.5 works for SH3/SH4 now

* Dont set PREFERRED_GCC_VERSION for SH

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

gcc-4.5: Fix toolchain builds for SH4/SH3

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

eglibc: Make it configurable by tuning eglibc options

* Look for disabled features in DISTRO_FEATURES and disable them while building eglibc
* By default all features are built
* One can disable a feature by specifying DISTRO_FEATURES += "noinet6 nobsd ...." e.g.

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

midori: add git version(fix some Segmentation fault)

Midori had a bug that made it segfault when a link was
  clicked, before the page was entirely loaded,rendering the
  browsing experience very difficult(it segfaulted continually
  under normal browsing activities)

Midori developers weren't sure what was the correct fix, but said that it
  was fixed in their git repository(from #midori on Freenode on irc):
    Jan 07 22:59:14 <kalikiana> GNUtoo|laptop, that looks like a bug that's fixed in git master
    [...]
    Jan 07 23:04:21 <GNUtoo|laptop> what's the bug number?
    Jan 07 23:04:33 <GNUtoo|laptop> so I could backport the fix's patch
    Jan 07 23:04:33 <nirik> I don't know off hand I'm afraid.
    Jan 07 23:04:35 <GNUtoo|laptop> ok
    Jan 07 23:04:39 <GNUtoo|laptop> what should I look for
    Jan 07 23:04:39 <GNUtoo|laptop> ?
    Jan 07 23:05:16 <nirik> possibly: http://git.xfce.org/apps/midori/commit/?id=1f15d17c5fcd6d3ce80a2f3bd5cd00780870fdf5
    Jan 07 23:05:32 <GNUtoo|laptop> thanks a lot
    Jan 07 23:05:37 <MonkeyofDoom> I thought http://git.xfce.org/apps/midori/commit/?id=97a7ae9955d4264f922452d6933f3a0e1e4b03f6 sounded suspect
    Jan 07 23:05:52 <MonkeyofDoom> but it's also just a possibility
    Jan 07 23:07:06 <kalikiana> the one MonkeyofDoom mentions is the crasher
    Jan 07 23:07:19 <kalikiana> the other one is a mis-behaviour of tab opening
    Jan 07 23:07:26 <GNUtoo|laptop> ok
    Jan 07 23:07:30 <GNUtoo|laptop> so many possibles crashes

Here's was the trace:
  Program received signal SIGSEGV, Segmentation fault.
  0x00000000 in ?? ()
  (gdb) bt
  #0  0x00000000 in ?? ()
  #1  0x401caef4 in g_object_unref () from /home/gnutoo/embedded/sshfs/usr/lib/libgobject-2.0.so.0
  #2  0x4116bd34 in webkit_hit_test_result_dispose(_GObject*) () from /home/gnutoo/embedded/sshfs/usr/lib/libwebkitgtk-1.0.so.0
  #3  0x401caef4 in g_object_unref () from /home/gnutoo/embedded/sshfs/usr/lib/libgobject-2.0.so.0
  #4  0x0006414c in midori_view_ensure_link_uri (view=0x22b008, x=0x0, y=0x0) at ../midori/midori-view.c:1630
  #5  0x00064210 in gtk_widget_button_press_event_cb (web_view=0x184058, event=0x229910, view=0x22b008) at ../midori/midori-view.c:1653
  #6  0x403ecf50 in _gtk_marshal_BOOLEAN__BOXED () from /home/gnutoo/embedded/sshfs/usr/lib/libgtk-x11-2.0.so.0
  #7  0x401ca538 in g_closure_invoke () from /home/gnutoo/embedded/sshfs/usr/lib/libgobject-2.0.so.0
  #8  0x401d2b88 in signal_emit_unlocked_R () from /home/gnutoo/embedded/sshfs/usr/lib/libgobject-2.0.so.0
  #9  0x401d6c04 in g_signal_emit_valist () from /home/gnutoo/embedded/sshfs/usr/lib/libgobject-2.0.so.0
  #10 0x401d6db8 in g_signal_emit () from /home/gnutoo/embedded/sshfs/usr/lib/libgobject-2.0.so.0
  #11 0x40500914 in gtk_widget_event_internal () from /home/gnutoo/embedded/sshfs/usr/lib/libgtk-x11-2.0.so.0
  #12 0x403eb590 in gtk_propagate_event () from /home/gnutoo/embedded/sshfs/usr/lib/libgtk-x11-2.0.so.0
  #13 0x403ebae0 in gtk_main_do_event () from /home/gnutoo/embedded/sshfs/usr/lib/libgtk-x11-2.0.so.0
  #14 0x40696e60 in gdk_event_dispatch () from /home/gnutoo/embedded/sshfs/usr/lib/libgdk-x11-2.0.so.0
  #15 0x40247940 in g_main_context_dispatch () from /home/gnutoo/embedded/sshfs/usr/lib/libglib-2.0.so.0
  #16 0x40247dfc in g_main_context_iterate.clone.4 () from /home/gnutoo/embedded/sshfs/usr/lib/libglib-2.0.so.0
  #17 0x40248260 in g_main_loop_run () from /home/gnutoo/embedded/sshfs/usr/lib/libglib-2.0.so.0
  #18 0x403ea8c8 in gtk_main () from /home/gnutoo/embedded/sshfs/usr/lib/libgtk-x11-2.0.so.0
  #19 0x000253cc in main (argc=1, argv=0xbeccece4) at ../midori/main.c:2202

The DEFAULT_PREFERENCE = "-1" is because of that(from #oe on FreeNode on irc):
  Jan 09 21:04:39 <Jay7> or commit with DP=-1 then post request to ML about promoting to default
  Jan 09 21:04:48 <GNUtoo|laptop> ok I'll send to ml
  Jan 09 21:05:14 <GNUtoo|laptop> I push some other stuff that only touch htcdream, and get thoses on ml
  Jan 09 21:05:15 <hrw> push with dp-1 first
  Jan 09 21:05:18 <GNUtoo|laptop> ok
  Jan 09 21:07:11 <GNUtoo|laptop> maybe I put DP -1 for all but not for SHR

In other words midori is used by a lot of distros/machines and as it's a git version,
  it's safer not to affect everyone in the case of a (more) problematic bug.

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