coreutils: Fix build with eglibc 2.16

eglibc 2.16 has removed gets so we account for that

(From OE-Core rev: 84183aee7e6f81919ec44f3043c7d9feea01f202)

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

glib: package bash-completion files in separate package and move ${PN}-codegen to .inc

* 2.32.4 has completion files in ${datadir}/bash-completion
  and dbus also packages bash-competion files in separate package

(From OE-Core rev: b070eb22372ad1ef5b21162d832772e560184cf8)

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

glib: upgrade to 2.32.4

* it's bugfix release, see http://ftp.gnome.org/pub/gnome/sources/glib/2.32/glib-2.32.4.news
  for list of fixed bugs

(From OE-Core rev: 85a930188300152bc9a1bbf640a2441f3177e26f)

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

libx11: fix makekeys build

* on hosts without libx11 installed it fails with
  | makekeys.c:33:27: fatal error: X11/Xresource.h: No such file or directory
  since oe-core rev 52d4ff7b35602faa52360a3e7a800891ef60c739
* upstream added -I$(top_srcdir)/include in 24283d40b1e4314c6647dda49d2a159833341a8b
  to ensure that right Xresource.h version is used, but we're building
  makekeys outside source tree (x11_disable_makekeys.patch), so we need
  to pass -I${S}/include in our makekeys build too
* building makekeys outside source tree isn't probably needed anymore
  and recipes should be updated to reflect that, upstream fixes:
  https://bugs.freedesktop.org/show_bug.cgi?id=22584
  http://cgit.freedesktop.org/xorg/lib/libX11/commit/src/util/Makefile.am?id=b2487d07f7b355f230a56e32c763acd08862735c
  http://cgit.freedesktop.org/xorg/lib/libX11/commit/src/util/Makefile.am?id=24283d40b1e4314c6647dda49d2a159833341a8b

(From OE-Core rev: abdc37d181019b9dd21ad08ad998ea4ff2a3e264)

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

busybox: Include sys/resource.h for rusage definition

Seen with eglibc 2.16

(From OE-Core rev: 077bd603040ebd16f01b4fa20999fb71a9a43695)

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

gdb: Replace struct siginfo with proper siginfo_t

Fixes errors like

/home/kraj/work/angstrom/sources/openembedded-core/build/tmp-eglibc/work/armv5te-oe-linux-gnueabi/gdb-7.4.1-r0.0/gdb-7.4.1/gdb/linux-nat.h:79:18:
error: field 'siginfo' has incomplete type

(From OE-Core rev: 6e5b37403dd4d0a9a249535d07b36bbae98cadd6)

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

mirrors: Add debian archive snapshot mirrors to search list

I ran across this problem when apt 0.7.14 was not fetchable
from regular debian mirrors and existing snapshot mirror
did not have it either since we did not use proper syntax
so it was not hitting that.

(From OE-Core rev: 5a5f67e5486837b4d7e5a89cc27a5c8f8c8598ef)

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

lsof: define linux C library type when using eglibc

lsof tries to compile a temp c source file and execute the binary to
determine linux C library type (file Configure, line 2689-2717).
It is inpracticable for cross-compilation and may have build issue on
some distros since it depends on host settings.

Fix below error when building for 64bit target on 64bit host:
[...]
| dsock.c:481:44: error: 'TCP_LISTEN' undeclared (first use in this function)
| dsock.c:482:45: error: 'TCP_CLOSING' undeclared (first use in this function)
[...]
| make: *** [dsock.o] Error 1

The actual issue exists in do_configure:
[...]
Testing C library type with cc ... done
Cannot determine C library type; assuming it is not glibc.

Which is in turn caused by missing 'gnu/stubs-32.h" when compiling
the temp c source file on host:
[...]
fatal error: gnu/stubs-32.h: No such file or directory compilation terminated.

file gnu/stubs-32.h is provided by 32bit glibc.

(From OE-Core rev: fbf22dfe1a0dca3fc7ac56e4fd274c0145efbc70)

Signed-off-by: Ting Liu <b28495@freescale.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base.bbclass: add GNUmakefile to the list of possible makefile names

(From OE-Core rev: b753eab1028ba48dfdcdeefa07f3f30743f3ee45)

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>

crosssdk/nativesdk: Ensure EXTRA_OECONF_FPU is unset

If EXTRA_OECONF_FPU is left set, certain ARM variables related to hard-float
can get pulled in and trigger rebuilds of the crosssdk code. The best solution
is to simply force the variable to a known correct value for the SDK targets
currently supported in the same way as TARGET_FPU.

There is some slight rearrangement of the gcc code to ensure the variable is
always used to call the fpu function.

(From OE-Core rev: 410990445ada8cdcfaec4e6fa5791cee9a5b8983)

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

python: Fix python-nativesdk variable contamination issue

Without this, python-nativesdk-crypt can depend on openssl which is incorrect.

(From OE-Core rev: 5740696c81a9143504c1cb93e94c54add27896e4)

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

sstate.bbclass: Move crosssdk before cross to ensure correct variables are set

There was a bug showing up where the crosssdk recipes were being installed as machine
specific in the sstate-control directory. This turned out to be due to the architecture
fields used by sstate being set incorrectly. The problem is that the crosssdk inherits
the cross class. It therefore needs to be listed in this if statement block before
the cross check, not after.

This should resolve some package architecture issue of crosssdk sstate files.

(From OE-Core rev: 79fe28e6033273f9632ca7549d1599d3fd1463ef)

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

gmp: upgrade to 5.0.5

(From OE-Core rev: 9c8000fdf696e234301a67d51dfeccc9bd1fcace)

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

sysvinit-inittab_2.88dsf.bb: Allow multiple serial port consoles to be defined

Set SERIAL_CONSOLES if you want to define multiple serial consoles, also if
you need to check for the presence of the serial consoles you can also define
SERIAL_CONSOLES_CHECK to determine if these are present when you boot. This
will prevent error message that pop up when the serial port is not present.

SERIAL_CONSOLES = "115200;ttyS0 115200;ttyS1 115200;ttyEHV0"
SERIAL_CONSOLES_CHECK = "${SERIAL_CONSOLES}"

The above lines in machine.conf or elsewhere will have the effect of having
two serial consoles and removing any that are not present at boot

(From OE-Core rev: eecd65f188ba2c924626b7e1bf5c8e2eb51e9b59)

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

runqemu: Fix running qemu when build without gl

When gl is disabled in PACKAGECONFIG then we dont need
to check for supporting libs to be present before running
qemu.

(From OE-Core rev: 6612873dc59054e6d37fa7488226218bfb759127)

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

image_types.bbclass: Default XZ_COMPRESSION_LEVEL to -e -6

Having XZ_COMPRESSION_LEVEL on -e -9 and -T 0 will make xz eat more
than 6Gb memory. Reduce this to -6 to make xz to use about 471Mb
on the tested machine.

(From OE-Core rev: cfd201ecda95d00bc5d70aa5045c649b98ffe8b8)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image_types.bbclass: Add XZ variable to set number of threads to be used while compressing

Default this variable to 0. This will make xz use as many threads as there are CPU
cores on the system.

(From OE-Core rev: 81c8e8a81e90ce95042e9fff8bac6fa011c416d1)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image_types.bbclass: Fix COMPRESS_CMD for xz to redirect compressed data to file

Having -c modifier makes xz to output the compressed data to stdout. In this
way the needed data will be in the do_rootfs log.
Redirect data to ${IMAGE_NAME}.rootfs.${type}.xz .

(From OE-Core rev: 7ceea61636d6a6f886002e3c29941da3356157ac)

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

cpan.bbclass: Fix CCFLAGS.

We should not just replace CCFLAGS with CFLAGS while compiling, because
this may cause run-time errors with perl's DynaLoader.pm.

Tested on qemux86 with new libnet-libidn-perl bb recipe:

root@qemux86:~# perl -e "use Net::LibIDN"
Not a CODE reference at /usr/lib/perl/5.14.2//DynaLoader.pm line 213.
END failed--call queue aborted at /usr/lib/perl/vendor_perl/5.14.2//Net/LibIDN.pm line 213.
Compilation failed in require at -e line 1.
BEGIN failed--compilation aborted at -e line 1.

(From OE-Core rev: 855211ae7a224f96663862d4a0e58a90d945dd48)

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

puzzles: convert to use SVN

Recently the tarball has been changing due the fact it gets rebuilt
nightly and timestamps are getting updated and changed internal to
the tarball.  Moving back to SVN to try and address this.

(From OE-Core rev: 8f30d41ede28a86564c402ce5b467e51aec6b15d)

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

Update proto -> protocol for recipes

(From OE-Core rev: 247fbed5a93800c44dc36e403f13bdc8dd704553)

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

xz: update to 5.1.2alpha

This contains bugfixes from the 5.0.3 and 5.0.4 releases

(From OE-Core rev: bbdbb70192e16fe351f9cf8f781110faf839be51)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

task-core-sdk.bb: add libgomp and libgomp-dev by RECOMMENDS

(From OE-Core rev: b2b682331cc1ec54448b5002ff94672d09fe9e7c)

Signed-off-by: Zhenhua Luo <b19537@freescale.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python-dbus: upgrade to 1.1.0

Tested by compiling core-image-basic

(From OE-Core rev: 39b2a455ab6b7c38b782e1b3e2811c3d9e4de7af)

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

classes/rootfs_deb: use more reliable check for package existence

It turns out "apt-cache showpkg" does return some information when a
package does not exist but another package recommends it, which can
occur for empty *-dev packages; so use "apt-cache policy" with a
different line count instead.

(From OE-Core rev: ac62761a9cacdfd0225d2db5a75584e6c85469a8)

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>

linux-libc-headers: updated to 3.4.3

Tested by building core-image-sato-sdk.

(From OE-Core rev: d67e47bd457de90b19428245644dc64ecf4ba11a)

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

package.bbclass: Allow overriding of debugedit starting path

In the usual case, the computed path used for debugedit to fix
up path references for the target filesystem is correct. However,
prebuilt binary components, such as prebuilt toolchains, can
have debug paths that do not reflect paths within the local
build directory. Providing an override lets us continue to use
the standard debugedit logic in package.bbclass.

(From OE-Core rev: 32b1fde106bf423873a4a56e1111f230494e2d4a)

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

runqemu/mips: adjust runqemu script to support mipsel machine

(From OE-Core rev: a12e2409af246d4c1d0e0c9a2457fb9c850bd8ca)

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

slang: change char type to signed char in macros

Fix [YOCTO 2726]

C language has 3 distinct char types:
         char
         unsigned char
         signed char
A char has the same range of values as signed char on X86,
but same as unsigned char on ARM which made Slang's typecast()
and array_sort() unable to work for char value on ARM, since
Slang is assuming "char" as "signed char".

Now we change "char" as "signed char" explicitly in
util/mkslarith2.sl, and use it to regenerate src/slarith2.inc

(From OE-Core rev: 0a1d842e2db63265a6a6708b1101e1f52f223fc9)

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

gconf: Disable gtk support

There are only a couple of helper utilities within gconf that need gtk+ as a
dependency and those are unused and pretty useless. We might as well drop
the dependency on gtk and allow more parallel builds by reducing dependency
bottlenecks.

(From OE-Core rev: 6facee8443966b646cd1e72f14ae13e58a13f621)

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

rpm_5.4.9.bb: fix builds on systems with older libc

| dbconvert.o: In function `rpmdb_convert':
| /local/jenkins/jobs/yocto-upstream/workspace/label/master/machine/p4080ds/poky/master/tmp/work/x86_64-linux/rpm-native-5.4.9-r45/rpm-5.4.9/tools/dbconvert.c:126: warning: the use of `tempnam' is dangerous, better use `mkstemp'
| dbconvert.o: In function `main':
| dbconvert.c:(.text+0x923): undefined reference to `htobe32'
| dbconvert.c:(.text+0xaa4): undefined reference to `htole32'
| dbconvert.c:(.text+0xac9): undefined reference to `htole32'

(From OE-Core rev: add2c772cd404e8bea4828959fcb5ee33c35c048)

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

mkelfimage: Upstream fixed packaging issue so upgrade to e1e6a91

Address QA issue: Files/directories were installed but not shipped.
New revision contains patch to mkelfimage removing
/usr/share/mkelfImage/elf32-i386 directory from Makefile install target.

http://review.coreboot.org/#/c/1195/

(From OE-Core rev: 7dd161816ccbb03db4f86902f96d3e86e09ccb64)

Signed-off-by: Raymond Danks <ray.danks@se-eng.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package install failure on systems without /sbin/ldconfig

Package install failures due to issues in post install rules.

package_do_shlibs only looks for libraries in a directory "lib", this should be
modified to the variable baselib.

ldconfig_postinst_fragment failure observed on systems without /sbin/ldconfig
        [ -x /sbin/ldconfig ] && /sbin/ldconfig
results in the post install rule returning a failure. Modify to
        if [ -x /sbin/ldconfig ]; then /sbin/ldconfig; fi

(From OE-Core rev: 3f85ab6d0d05a4e34fe73fa343d2b3b942fc69b7)

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

conf/machine: replace TUNE_CONFLICTS with TUNECONFLICTS

* it wasn't consistent with other machine configs
* reported 2 months ago..
  http://lists.linuxtogo.org/pipermail/openembedded-core/2012-May/022154.html

(From OE-Core rev: 3fec966531059b4b21f40be3b22a60edf88c5190)

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

perf: add perl, python to DEPENDS

perf has perl and python compile-time dependendencies, add them.

(From OE-Core rev: 1b40dac2c501ae8c8d812cf2866fe7917f144a19)

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

linux-yocto/3.0: perf scripting support

Updating the 3.0 SRCREVs to pickup the recent perf scripting fixes.

  cef17a1 perf: add sgidefs.h to for mips builds
  3258da3 perf: change --root to --prefix for python install
  1f9c25b perf: add 'libperl not found' warning
  f2b1bf1 perf: use pkg-config instead of python-config
  568e05f perf tools: Add support to install perf python extension
  7e47f81 Revert "perf: hard-code NO_LIBPERL/NO_LIBPYTHON"

(From OE-Core rev: 20405bdab18b7a2b101f818e72863d289d5945e4)

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

bitbake: fetch2/__init__.py: Warn user if SRC_URI is using "proto" and not "protocol"

As well, if "proto" is used, get the associated value as "protocol"

(Bitbake rev: 53e6b630f0463d2d07cdaa9c9eb36794dc9b6b69)

Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: fetch2/bzr.py: Use "protocol" parameter to get the protocol

(Bitbake rev: 630876b40ed181312e84f902c4cfb97361afbe81)

Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: fetch2/hg.py: Use "protocol" parameter to get the protocol

(Bitbake rev: f4a780a7f50fdd8f2bd75888dad790bcfe95b873)

Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: fetch2/osc.py: Use "protocol" parameter to get the protocol

(Bitbake rev: 01040efce82cfbaa76f46d9f95c984be8f9ea810)

Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake: Abort build if runtime dependency conflict

Currently if there are multiple preferred providers available for
a runtime dependency, bitbake will print an Error message and let
the build go on. Anyways the build should abort while any Errors
occured.

[YOCTO #2734]

(Bitbake rev: 5f81a714f4fca785780bef555b419f0250e5ec1c)

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

connman-gnome: remove unused patches, they were all merged upstream

(From OE-Core rev: bf134eaf752cd5db6cce79ad4c1ae0f25d46b08e)

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

libgomp: add libgomp (openmp) library, and build for powerpc targets by default

(From OE-Core rev: b76a7bc8dcb7aed7d6f026e77a226837004c50af)

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

gcc: gcc-cross-canadian: use correct location for libraries for powerpc64

This fixes the issue where gcc invokes the linker with an incorrect -L
library location and gives up because it can't find libraries. It was
looking in a /lib folder instead of /lib64

(From OE-Core rev: 1b428f0151a5b86332b501871bf4a925b43580a5)

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

gcc-configure-common.inc: use --with-long-double-128 on powerpc to comply with ABI

(From OE-Core rev: 243539efec94e79dade2525d0ab3586eea48c529)

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

openjade-native_1.3.2.bb: fix typo and change the deps exclusion to correct var

(From OE-Core rev: 4be17f018f8dd0ed576498badc03ee13095b2263)

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

dtc.inc: fix for libdir == /usr/lib64

On 64bit systems dtc will still install libaries in /usr/lib
unless we havet this override

(From OE-Core rev: b60579e66601ba60669a9e1194409fcd8b80530e)

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

valgrind: fix default.supp missing issue

    When run valgrind, following error appears:
    ==2254== FATAL: can't open suppressions file "/usr/lib/valgrind/default.supp"

(From OE-Core rev: 0b3261d513cdad80174a9b9e804981c50bcb7ca2)

Signed-off-by: Zhenhua Luo <b19537@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

directfb: disable imlib2 explicitly

When HOST machine installed imlib2, compiling directfb in OE will
direct this library, and try to use it. This is wrong, since there is
no imlib in OE, so we should disable it explicitly here.

Error as following:
| idirectfbimageprovider_imlib2.c:64:20: fatal error: Imlib2.h: No such
file or directory

Config.log as following (which is wrong):
configure:19602: checking for imlib2-config
configure:19620: found /usr/bin/imlib2-config
configure:19633: result: /usr/bin/imlib2-config

(From OE-Core rev: fbdee213465f38791f3dd6c2663f2fe295be3596)

Signed-off-by: Dennis Lan <dennis.yxun@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

subversion: do_install failed (parallel issue)

install-neon-lib needs libsvn_delta-1.la which will be regenerated
during libsvn_delta-1.la's installation, if libsvn_delta-1.la is
in regenerating and at the same time install-neon-lib links it, the
error willl happen.

The error message is:
/bin/ld: cannot find -lsvn_delta-1
collect2: error: ld returned 1 exit status

This is a parallel issue, so it doesn't happen often.

Note:
The autoreconf doesn't generate build-outputs.mk, it would be generated
by autogen.sh (use build.conf as the input), but autogen.sh isn't
suitable for cross compiling, so both modified build-outputs.mk and
build.conf.

[YOCTO #2727]

(From OE-Core rev: ce37c45abb4cf43e5009867f695982de2eb33450)

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

gnupg: ensure deterministic build by disabling libcurl

gnupg has it's own fake curl, since we use gnupg with zypper, there does not
seem to be a strong reason to add curl to the depends list.

(From OE-Core rev: 44ea2cf548ca20488449d347f8f3b7d49f7e7493)

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

perf: add perf-tui feature

Add a new feature named 'perf-tui'.  Adding this into the
PERF_FEATURES variable in perf.inc will enable the perf TUI (Text-base
UI) user interface on a target, which adds libnewt and turns on the
perf text UI options in perf, if perf is included in an image.

If 'perf-tui' isn't named as a feature (the default), the perf TUI
will be disabled and unavailable.

(From OE-Core rev: 4fd8a550886f02189e4ed127d0a2f16e92f8474c)

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

perf: add perf-scripting feature

Add a new feature named 'perf-scripting'.  Adding this into the
PERF_FEATURES variable in perf.inc will enable perf scripting on a
target, which will turn on all the language bindings currently
available in perf (Perl and Python), if perf is included in an image.

If 'perf-scripting' isn't named as a feature (the default), all perf
language bindings will be disabled and unavailable.

(From OE-Core rev: fc0661041436013b9099dbd659994a2b8b292c19)

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

perf: add perf.inc

Add a perf.inc to contain utility functions and definitions and to
avoid cluttering up the main recipe.

(From OE-Core rev: 414d00be5b350ea84fc7e1ff690f78b3396cfcd0)

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

linux-yocto/3.4: perf, unionfs and yaffs2 fixes

Updating the 3.4 SRCREVs to pickup the following changes:

  6297e4c perf: add sgidefs.h to for mips builds
  69b7817 yaffs: add leading underscore to mtd_info function names
  160f1ac yaffs: change type for mode to umode_t
  2d875e8 Unionfs: use mode_t

(From OE-Core rev: 9d75c34948f5d961cff9d72fbaa628ff58821cc7)

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

linux-yocto/3.2: perf: fix mips build

Updating SRCREVs for the following fix:

Author: Tom Zanussi <tom.zanussi@intel.com>
Date:   Tue Jul 10 09:38:40 2012 -0500

    perf: add sgidefs.h to for mips builds

    Allow Python.h to find the definitions it needs on mips i.e. get rid
    of the error: "_ABIN32" is not defined.

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
(From OE-Core rev: 04396899dc7973261404d5fbec3f85aa03a32e03)

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

opkg: fix installation of dev / dbg packages

OE-Core revision ffe93e2a099e4cadb1829dc12a58a6d7bef9a5a7 moved opkg's
configuration file into /etc/opkg, and opkg always reads all files under
this directory even if -f is specified (as it is during do_rootfs in
order to specify our own generated config file), and this means that
after opkg-collateral is installed into the rootfs, this version of the
config is used, resulting in opkg ignoring the list of packages it has
already downloaded (since it now sees list_dir as /var/lib/opkg instead
of the default /var/lib/opkg/lists) and thus it fails to install any of
the "attemptonly" packages (including *-dev, *-dbg, etc.).

If we change the -f option to ignore configuration files in /etc/opkg
then we no longer have this problem.

Fixes [YOCTO #2595].

(From OE-Core rev: 37df134557802ba116f001597b7cd5e9bc39e188)

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

bitbake: fetch2/wget: Use FETCHCMD_wget to provide the commandline and options

This also changes to use  -t 2 -T 30 as the defaults which are more sane for
a modern fetcher and already specified in OpenEmbedded metadata.

(Bitbake rev: bf0e5dddf0f63cdb0648fb6d872af5ceef6fbfb0)

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

bitbake: fetch2/cvs.py: Switch to use FETCHCMD_cvs and handle parameters in the fetcher itself

This brings the cvs fetcher more into line with the others and allows consistent usage
of the FETCHCMD variable and option handling.

(Bitbake rev: 390ad59739356422852e976fa246375abc6aba08)

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

ccache: Ensure CCACHE_DIR is cleaned by do_clean task

This ensures a rebuild does really build from scratch when ccache is in use.

[YOCTO #2696]

(From OE-Core rev: ddf52d0161096b089fad8f3ace69b6515d7b7226)

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

libxml2: Fix libzypp ansidecl related build failures

cmake looks at all include statements, even if they're not used. To make
builds deterministic and avoid needing to add binutils as a dependency
for libzypp, completely remove the include from the header file, even if
it is never used.

This avoids issues where you'd build binutils, then libzypp, then remove
binutils (and hence ansidecl.h) and then recompile libzypp which would
still have the dependency and hence fail.

(From OE-Core rev: bfaaeb44c5023e2d2a9414c07694c75fa527283b)

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

bitbake.conf: Drop the remainder for the FETCH/CHECK/RESUMECOMMAND variables

This completes the removal of these older variables and their replacement
with FETCHCMD. This change requires the latest bitbake to operate optimally.

(From OE-Core rev: 7dd2a12ebe49865720dcea60cd2ed13205c0c4fd)

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

bitbake.conf: Add passive-ftp option to the default FETCHCMD_wget

This brings it into sync with the wget FETCHCOMMAND and RESUMECOMMAND
variables.

(From OE-Core rev: 0f8290ff1c9bc86c744f67ce6194dd6911533012)

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

bitbake.conf: Drop the old svn FETCH/UPDATE/RESUMECOMMAND variables

The svn fetcher converted to use FETCHCMD a long time ago. This drops
several variables which are effectively useless.

(From OE-Core rev: 34a8cdafee145d04f8b526b094341c34cf36ba1b)

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

core-image-*.bb: Remove usage of EXTRA_IMAGE_FEATURES from some .bb files

Replace the usage of EXTRA_IMAGE_FEATURES in a small number of .bb
files with IMAGE_FEATURES, and leave the use of EXTRA_IMAGE_FEATURES
for developers in their local.conf files, to avoid the possibility of
undesirable side effects.

(From OE-Core rev: 347fc6e82a3d1c1ac7dbabd9e5cdd08abd864bfb)

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

core-image-*: remove copyright notices

These image recipes are meant to be examples that people can copy and
modify completely for their own purposes, and most of them are so
trivial they don't really need copyright notices anyway, so trim them
off.

(From OE-Core rev: 22db8917f6301595a48f6dcbcccc27627c8a3941)

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

core-image-minimal-dev: base on core-image-minimal

When core-image-minimal has been updated in the past, we have on several
occasions missed applying those updates to this recipe, so let's just
base it directly on core-image-minimal and completely avoid that problem
in future.

(From OE-Core rev: 7681e50da86606f4d7b63504f0c9233a1c87b642)

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

libgsmd - remove

This project has been unmaintained for some time, and even the OpenMoko
project is not using it any more (in favour of FSO). Since we have ofono
in OE-Core which replaces and surpasses its functionality, we can remove
libgsmd.

(From OE-Core rev: 4cd880c61e9d74dbf1a747f3654239cadadf45ce)

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

linux-yocto/3.2: perf scripting updates

Updating the 3.2 SRCREVs to reflect the perf scripting fixes by TomZ

standard/default/base:
  0ec416e perf: change --root to --prefix for python install
  18ad076 perf: add 'libperl not found' warning
  e0f641d perf: use pkg-config instead of python-config
  3195098 Revert "perf: hard-code NO_LIBPERL/NO_LIBPYTHON"

meta:
  07ee09b features/ftrace: enable FTRACE_SYSCALLS

(From OE-Core rev: c8f60299424dd912c6402847773f89c0490a1857)

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

bitbake: bitbake/cooker: Print which pkgs would be built in -g output

This is for giving the user a clear list to show which pkg would be
built, we have the "bitbake -g", but it is not easy to read for people,
it is for "dot". Improve the "bitbake -g" to also save a pn-buildlist:

$ bitbake -g core-image-sato
...
NOTE: PN build list saved to 'pn-buildlist'
[snip]

The contents of pn-buildlist:

busybox
shadow-native
pth
sysfsutils
qemu-helper-native
curl-native
ncurses-native
gdbm
xserver-xorg
linux-libc-headers
[snip]

[YOCTO #2404]

(Bitbake rev: 18aff925aece774d0172894e25584353519ca03f)

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

linux-yocto/3.4: perf scripting & unionfs fixes

Updating the SRCREVs to pickup the following changes:

standard/base:
 95c79fb Unionfs: update ->show_options prototype
 49f1599 perf: change --root to --prefix for python install
 8ff1ab6 perf: add 'libperl not found' warning
 b4bc865 perf: use pkg-config instead of python-config
 a2257bf Revert "perf: hard-code NO_LIBPERL/NO_LIBPYTHON"

meta:
 a8cf770 features/ftrace: enable FTRACE_SYSCALLS

(From OE-Core rev: 8d8b9e77aa3403e880cbbdfefc7f24a993b3161f)

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

psplash - Update to latest psplash version. LIC_CHKSUM Tweak

* Update psplash to the latest version.
* Change the license checksum to use the lines in the psplash.h that contains
  license information instead of doing a checksum on the entire file.

(From OE-Core rev: e15dd9dede74c8906e118647ab93941c1fcb35b1)

Signed-off-by: Franklin S Cooper Jr <fcooper@ti.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel.bbclass: add non-santized kernel provides

If the kernel version string uses characters or symbols that
need to be santized for the package name, we can end up with a
mismatch between module requirements and what the kernel
provides.

The kernel version is pulled from utsrelease.h, which contains
the exact string that was passed to the kernel build, not
one that is santized, this can result in:

 echo "CONFIG_LOCALVERSION="\"MYVER+snapshot_standard\" >> ${B}/.config

 <build>

 % rpm -qp kernel-module-uvesafb-3.4-r0.qemux86.rpm --requires
update-modules
kernel-3.4.3-MYVER+snapshot_standard
 % rpm -qp kernel-3.4.3-myver+snapshot-standard-3.4-r0.qemux86.rpm --provides
kernel-3.4.3-myver+snapshot-standard = 3.4-r0

At rootfs assembly time, we'll have a dependency issue with the kernel
providing the santizied string and the modules requiring the utsrelease.h
string.

To not break existing use cases, we can add a second provides to the
kernel packaging with the unsantized version string, and allowing the
kernel module packaging to be unchanged.

   RPROVIDES_kernel-base += "kernel-${KERNEL_VERSION}"

 % rpm -qp kernel-3.4.3-myver+snapshot-standard-3.4-r0.qemux86.rpm --provides
kernel-3.4.3-MYVER+snapshot_standard
kernel-3.4.3-myver+snapshot-standard = 3.4-r0

(From OE-Core rev: 7be043178f5b9d16d9a06696bc0b96689f202a8d)

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

connman: ship the empty plugins directory in the connman package

Otherwise we get a QA warning that /usr/lib/connman/packages isn't shipped.

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

connman: remove some obviously wrong (circa early-GNOME 2?) directories from FILES

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

ncurses: fix packaging issue in multilib build

ERROR: QA Issue: lib32-ncurses: Files/directories were installed but not shipped
  /usr/bin/clear.ncurses
  /usr/bin/reset.ncurses

(From OE-Core rev: e3a29405a59df45ff72f088c602181e1435d8abe)

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

bluez4: fix packaging issue after update

WARNING: QA Issue: bluez4: Files/directories were installed but not shipped
  /usr/share/dbus-1
  /usr/share/dbus-1/system-services
  /usr/share/dbus-1/system-services/org.bluez.service

(From OE-Core rev: 8635372d8e1886d3f0c066341f6909e5015d81f6)

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

packagedata.py: Fix get_subpkgedata_fn for multilib

This happens when tryng to add libgcc-dev to as a multilib package
(e.g. IMAGE_INSTALL_append = " lib32-libgcc-dev")

| Processing task-core-boot...
| Processing fman-ucode...
| Processing dosfstools...
| Processing lib32-libgcc-dev...
| Unable to find package lib32-libgcc-dev (libgcc-dev)!
NOTE: package fsl-image-full-1.0-r1.1.3.6: task do_rootfs: Failed

RPM (or bitbake?) is looking in the tmp/pkgdata, however some of these file
paths are mungned for the multilib scenario:

$ find tmp/pkgdata/ | grep libgcc-dev$
tmp/pkgdata/ppce5500-fsl-linux/runtime/lib32-libgcc-dev
tmp/pkgdata/ppc64e5500-fsl-linux/runtime/libgcc-dev

This patch fixes where we look for these files so they can be found and
properly installed for the multilib root file system

(From OE-Core rev: 4e65ddda5d351bd9f2172e77e656903f61a52f34)

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

qemu-0.15.1: add patch to fix compilatation problems on powerpc

ERROR: Function failed: do_compile (see /opt/yocto/cache-build/p5020ds-64b/build_p5020ds-64b_release/tmp/work/ppc64e5500-fsl-linux/qemu-0.15.1-r6/temp/log.do_compile.28447 for further information)
ERROR: Logfile of failure stored in: /opt/yocto/cache-build/p5020ds-64b/build_p5020ds-64b_release/tmp/work/ppc64e5500-fsl-linux/qemu-0.15.1-r6/temp/log.do_compile.28447
Log data follows:
| DEBUG: SITE files ['endian-big', 'bit-64', 'powerpc-common', 'common-linux', 'common-glibc', 'powerpc-linux', 'powerpc64-linux', 'common']
| ERROR: Function failed: do_compile (see /opt/yocto/cache-build/p5020ds-64b/build_p5020ds-64b_release/tmp/work/ppc64e5500-fsl-linux/qemu-0.15.1-r6/temp/log.do_compile.28447 for further information)
| NOTE: make -j 24
|   LINK  ppc-linux-user/qemu-ppc
| /opt/yocto/cache-build/p5020ds-64b/build_p5020ds-64b_release/tmp/sysroots/x86_64-linux/usr/libexec/ppc64e5500-fsl-linux/gcc/powerpc64-fsl-linux/4.6.4/ld:/opt/yocto/cache-build/p5020ds-64b/build_p5020ds-64b_release/tmp/work/ppc64e5500-fsl-linux/qemu-0.15.1-r6/qemu-0.15.1/ppc64.ld:84: syntax error
| collect2: ld returned 1 exit status
| make[1]: *** [qemu-ppc] Error 1
| make: *** [subdir-ppc-linux-user] Error 2
| make: *** Waiting for unfinished jobs....
| ERROR: oe_runmake failed

(From OE-Core rev: a9207aad5b163a071cd8298517d61514c587e0ed)

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

libxml-parser-perl_2.41.bb: fix MakeMaker issues with using wrong CC/LD/etc

MakeMaker has a bug where it does not propagate CC/LD/etc information
down to subproject it generates Makefiles for... this recipe has has an
Expat subproject which has issues building if we are using sstate-cache
and it will reference the old sysroots and be unable to build properly.
There is an upstream MakeMaker bug for this issue but we can work around
it by fixing up the Makefiles for now

See:
https://rt.cpan.org/Public/Bug/Display.html?id=28632

(From OE-Core rev: 2c3629a698e537c69c71039bc6172b3c0e617ff1)

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

utils.bbclass: add helper function to add all multilib variants of a specific package

This is useful for the scenario where we want to add 'gcc' to
the root file system for all multilib variants

(From OE-Core rev: 7cd54693363882854cd026ebe071e9f2e03d364e)

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

pseudo: Update to 1.3.1 (fixing chroot crash)

Yocto bug #2639.  If a chroot path was long, expanding absolute
paths within the chroot path could overrun a buffer.

(From OE-Core rev: 282cbf7002112f6b58ca1ee070c0b1285b838db7)

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

grub-efi: Do not use help2man

Fixes [YOCTO #2527]

Modify configure.ac and the generated configure script to avoid using
help2man during the compilation process. For grub-efi we are only
deploying the EFI payload and are not installing grub on the target
root filesystem. Therefor, we do not need the man pages.

Cleanup the SRC_URI whitespace while we add a line to it.

(From OE-Core rev: 035bc3f5ee99ed3bd74219d9717239299e4bc765)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
CC: Paul Eggleton <paul.eggleton@linux.intel.com>
CC: Radu Moisan <radu.moisan@intel.com>
grub-efi-native whitespace cleanup (INC)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

init-install: Clean up partition alignment

The current partitioning scheme leaves a 1MB gap between all the
generated partitions by adding a 1 to the end of the last partition to
use as the start of the next. parted is smart enough to not overlap
start and end positions of the same value. This avoids the 1 MB gaps.

Rather than pad the disk with 1MB in the beginning and cut it off at the
MB boundary on the end, we can use 0% and 100% to allow parted to do the
required math and use as much of the disk as possible.

(From OE-Core rev: 8aac6ecc5194c734dfd3d677017ab3ea045b2339)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

init-install: Correct ext2->ext3 typo in logging

We create both the boot and root partitions as ext3 now, update the
logging accordingly.

(From OE-Core rev: 4436639eed57d818992596d6f0f7b53d3bbd4800)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

init-install: Use swap_ratio in the calulation of swap_size

swap_size currently uses a hard coded percentage and ignores the
swap_ratio variable. Fortunately they are the same value currently. Make
the calculation use the variable to avoid problems in the future.

(From OE-Core rev: 2678ce668499af0e90994b9da8c518e85de56651)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/license: fix manifest to work with deb

Prepend the license manifest creation call to ROOTFS_POSTPROCESS_COMMAND
instead of appending to ROOTFS_POSTINSTALL_COMMAND. The latter is not
implemented for the deb backend (and probably ought to just be removed
completely), and by using _prepend we can still ensure it occurs before
package info is removed (and before buildhistory in case it is needed
there in future).

(From OE-Core rev: 56fafe083fdb63f156a081344fb1d836e3182bf0)

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

runqemu: fix support for ext4 rootfs images

(From OE-Core rev: c9479ae46d0b891eda8f0db89bc66fdf08c3f7c2)

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

runqemu: fix usage() help for MACHINE setting

(From OE-Core rev: 36482a0064993b047829631d063beadbc03f2cbf)

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

sanity.bbclass: the tmpdir can't be longer than 410

There will be errors when the length of the tmpdir is longer than 410:

1) Longer than 420:
Can't exec "/bin/sh": Argument list too long at /usr/lib/perl/5.10/IO/File.pm line 66.

This error happens on both Ubuntu 10.04 and 10.10 when the pkg needs run
"autoreconf", this is because it passes many files with absolute path to
aclocal, aclocal passes them to perl, this is a limitation of the perl
on Ubuntu 10.04 and 10.10, and the perl-native is not ready at this very
early stage.

2) Longer than 490:
bitbake/lib/bb/persist_data.py", line 197, in connect(database=...)
    >    return sqlite3.connect(database, timeout=5, isolation_level=None)
OperationalError: unable to open database file

This error happens on Ubuntu 10.04, 10.10 and Fedora 17. This is because
the length of the database in sqlite3 module (host's) can't be longer
than 490 (or little smaller). The python-native is not ready at this
very early stage.

The 2 errors are host related, I think that limit the length of the
TMPDIR to 410 is OK for most of the build, rarely build sets TMPDIR's
longer than 410.

[YOCTO #2434]

(From OE-Core rev: ebcf949853ff667478a1ea1d3f1f8f41d643e708)

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

apt 0.7.14: runtime error: Method file has died unexpectedly

When the length of the tmpdir is longer than 400, there is an error
when run "apt-get update":

Method file has died unexpectedly!

This is because the "char S[1024]" is not enough for long URI, S[2048]
would be enough.

[YOCTO #2689]

(From OE-Core rev: 3ed08bd24cef39a85c528159a494e8f0dd001739)

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

apt 0.7.14: runtime error: filename too long (tmpdir length)

when the tmpdir dir is longer than 220, there is no files saved in
tmp/sysroots/x86_64-linux/var/lib/apt/lists/ after run apt-get update,
this is because apt-get uses the path as the file name, but the file
name can't be longer than 255 according to /usr/include/linux/limits.h.

[YOCTO #2688]

(From OE-Core rev: 9a0c0393871eda4bbcecfdd4b595f0c1b8e42edf)

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

curl: upgrade to 7.26.0

Config system changed from 7.24.0 and the noldlibpatch
is no longer needed, thus deleted.

(From OE-Core rev: 0d2d59420b5924491ccd5c091c823b9c277a6721)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-common: Don't use "is" for comparing strings, use "=="

Needed because the equality check was failing here even though upon
printing the LHS and RHS were the same.
As per http://stackoverflow.com/a/2987975/64537, using "is" compares the
memory addresses of the two objects which is not what we want here. We
just want to compare the values.

(From OE-Core rev: 7fd82cc90a48302ed42b6bfa962bb0de2c652b45)

Signed-off-by: Kartik Mohta <kartikmohta@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

recipes.txt: Add entries for "recipes-rt" and "recipes-support"

(From OE-Core rev: 9705ccc10892b15fa3269ed071bb41d45987694c)

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pixman: upgrade to 0.26.2

Also, disable Loongson MMI for now because Pixman's build system will
autodetect a Loongson2f architecture revision and the build will fail.

(From OE-Core rev: 7dded4da52bc73677583da740620b0c81e594d62)

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

xinput: upgrade to 1.6.0

Licence file changed, an entire part has been removed.

See diff below:

<<
 TORTIOUS  ACTION, ARISING    OUT OF OR   IN  CONNECTION  WITH THE USE OR
 PERFORMANCE OF THIS SOFTWARE.

-
-Copyright 2007 Peter Hutterer
-
-Permission to use, copy, modify, distribute, and sell this software and its
-documentation for any purpose is hereby granted without fee, provided that
-the above copyright notice appear in all copies and that both that
-copyright notice and this permission notice appear in supporting
-documentation.
-
-The above copyright notice and this permission notice shall be included
-in all copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
-OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
-MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
-IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY CLAIM, DAMAGES OR
-OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
-ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
-OTHER DEALINGS IN THE SOFTWARE.
-
-Except as contained in this notice, the name of the author shall
-not be used in advertising or otherwise to promote the sale, use or
-other dealings in this Software without prior written authorization
-from the author.
-
-
+Copyright © 2007 Peter Hutterer
Copyright © 2009 Red Hat, Inc.

Permission is hereby granted, free of charge, to any person obtaining a
>>

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

libxi: upgrade to 1.6.1

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

libx11-diet: upgrade to 1.5.0

The X18NCMSstubs.diff was modified to please the -Werror=return-type
compiler option. Otherwise, it will throw an error.

(From OE-Core rev: 708785266d1d671f5d2a8e8af81a77e34c70662f)

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

libx11-trim: upgrade to 1.5.0

(From OE-Core rev: 817742fa1a5261b3b2c697de3c14e602d55713e2)

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