sqlite3: Fix nativesdk packaging/QA warnings

When building sqlite3-nativesdk, there was a warning about debug files in the
main package. This was due to the order of items in PACKAGES with -dbg after
the main package which breaks an assumption native.bbclass makes. Changing
the order fixes the packaging problem with no change to the normal target
packaging.

(From OE-Core rev: 4f5fdc4dc14d287d301069024ddec9cb65d68f7f)

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

grub: Fix insane/QA architecture warning

There is QA warning about this package for an architecture mismatch but
this is inappropriate in this case since the bootloader needs 32 and 64
bit code. We therefore flag the QA check to be skipped.

(From OE-Core rev: 43723e19eb5a6119c7546dc812428e792999a928)

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

gcc: Fix two QA issues

a) There is a QA warning from a .so being present in a main package.
   In the case of the plugin library for gcc, this is allowed.
b) Remove the unwanted libiberty.a file with the strange path. We
   don't need/want this and this removes an unpackaged file warning.

(From OE-Core rev: ca36a3edf9cede9fa0d73ba1a9538ab467cb5e3c)

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

linux-yocto/meta: eg20t and live boot config changes

Merging the following configuration changes:

 67a46a6 meta/common-pc-64: enable live booting for common-pc-64
 1010905 meta/common-pc: enable live booting for common-pc
 b3c5fa7 meta/atom-pc: enable live booting for atom-pc
 41c090e meta: update boot live config and move it to cfg/
 d51b0e7 eg20t: update config options

The first 4 make the live-boot configuration shared and then reuse
them for the boards that currently are live bootable. The eg20t
is a cleanup of obselete kernel options and is part of the cleanup
of options for the 3.0 kernel.

[YOCTO: #940]
[YOCTO: #686]

(From OE-Core rev: da496bb570d05d0906a7310697bdd636c254540c)

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

avahi: fix avahi-systemd dependencies

During the avahi-ui split the RDEPENDS line got dropped leading to: http://git.angstrom-distribution.org/cgi-bin/cgit.cgi/testlab/commit/?h=yocto&id=ae0db0eb51bde42f09b9513842c3b745c37c60d5

(From OE-Core rev: b447c07168405b748d4db0641468590c10a7ad52)

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

blktrace: update to version 1.0.3

This also removes dso_linking_change_build_fix.patch since the same
change has been made upstream.

Also removes unused SRC_URI checksums.

(From OE-Core rev: 776942858cad072facc9b5fea618ee88ff56a3a8)

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

sysprof: update to version 1.1.8

Also removes unused SCR_URI checksums.

(From OE-Core rev: 78d4a8d095faa3915e0a53ee2fb4afe6429928c2)

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

libffi: really populate -dev package

As per gcc PR 11147, libffi installs headers into a target dependent
place (/usr/lib/....). Include a rule to include those files into
libffi-dev package.

Reference: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11147

(From OE-Core rev: 684a4b517d13884c315688967fadd5e6a4845b71)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gconf-dbus: packaging fixup

Behave more like plain gconf: include a dtd and .la files into -dev
package.

(From OE-Core rev: 9e962c1b4c8e5a3352f5e2b7dc162aeac1335b3a)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: include libgcov.a into libgcc-dev package

First, this lib is usefull for coverage analysis-enabled building.
Second, this fixes the warning about unpackaged files in libgcc recipe.

(From OE-Core rev: 2a807a98d8be3f486e703321773db32657c71d9e)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: fix possible problems with nscd compilation during eglibc-nativesdk build

Long time ago a066e7ca90a28d5681c5fa895a29e999ed7c88b was committed to
address possible problems with compilation of nscd during
eglibc-nativesdk build. Problems were related to the way gcc searches
for headers to check if it should enable it's own stack smash protection
bits or it can relay on eglibc for it.

However after 934d38530c9a67562e53d4034aee5531f0f26750 things got
broken, as for gcc-crosssdk-intermediate packages:
1) EXTRA_OECONF is ignored
2) headers are installed in a different location than expected by that
patch.

This results in eglibc-nativesdk build broken on some systems (e.g. mine
Debian x86_64 squeeze). Fix that by providing with-headers options to
crosssdk-intermediate gcc configuration.

(From OE-Core rev: 63494d638b7a9b88a5b7d7a02d2afcb3aa0fa064)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

icecc-create-env-native: provide the script right in the tree

There is no point in downloading a tarball with no clear upstream (other than
icecc itself) and then patching it. Rather put new script in the source tree.

(From OE-Core rev: 409fa8ca4d37ad407faaa2a8935e9d2bb89776c9)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake.conf: change ccache path to use MULTIMACH_HOST_SYS

Currently if I build packages for several targets (e.g. for armv5te tosa
and for armv7a beagleboard) oe will use single ccache dir for both of
those targets: build/ccache/arm-oe-linux-gnueabi. However those targets
use different opcodes, different features and binaries created for one
of those targets wont't run on the lower one. So use MULTIMACH_HOST_SYS
for ccache dir, so that it uses something like
build/ccache/armv5te-oe-linux-gnueabi dir.

(From OE-Core rev: 982373006a98cf2303514badd1cfb692108408c0)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnome-doc-utils: Prepend PKG_CONFIG_SYSROOT_DIR to the path returned from PKG_CONFIG

If we build say gnome based image on a build system which does not have
gnome e.g. kubuntu then packages like gedit do not build since it uses
gnome files from host system which are non existent on kubuntu

(From OE-Core rev: 7b18b3c96634e40abf690a7ec72562389b0e6c23)

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

libtool: Fix an issue where unnecessary rpaths were being injected

libtool was being to aggressive in adding rpath to binaries. This
change stops it adding them if that path the default search path.

(From OE-Core rev: 99d1e3ee56c326b3ab68913e68d133a63f334696)

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

sstate: No need to spew out a debug message per file, summarise instead

(From OE-Core rev: c7b02c6e80819e30a0818282ab8d960243a2d0e8)

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

base.bbclass: Drop legacy code that is no longer useful/used

The code displaying console events has been handled by the bitbake
UI since 1.8 so the legacy code path can be removed. If a log event
is wanted, there are much better (and more complete) ways to do this
so remove the EVENTLOG code too.

(From OE-Core rev: aa60484384385a8b0f07e8062d58056a3670e399)

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

autotools.bbclass: Set the dynamic linker search path for libtool correctly

libtool obtains the search path from /etc/ld.so.conf and hardcodes /usr/lib
and /lib. This results in host contamination and variable sets of RPATH
values ending up in binaries.

By exporting the correct values for all autotools recipes we avoid this.

(From OE-Core rev: 93e595d5c89ebacdb8d1e6fcfe6f58fe2d30de28)

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

matchbox: Upgrade SRCREV to reflect recent accpeted patches by upstream

(From OE-Core rev: 33a1a05ef988c69f8ff8e38c6723922082e5d1aa)

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

tcp-wrappers: Bump PR to make meta-oe merge

meta-oe has a version of tcp-wrappers 7.6 recipe
which has virtually no differences with the one from
oe-core. So with this patch we can remove the recipe
from meta-oe

(From OE-Core rev: 7c503d056d15e376243928ffb815296e9c711658)

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

neon: Add libproxy to DEPENDS to ensure determinstic builds

(From OE-Core rev: ed2a606909b9490ac57a3ad3db7a15e83a8664f9)

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

libsndfile1: Disable external codec librbaries since we don't list in DEPENDS

(From OE-Core rev: 34a14ce3ea78be299175e1a803f92519aa02355b)

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

cairo: Disable bfd symbol loopup since we don't list it in DEPENDS

(From OE-Core rev: ac8b7e275a8789b6605ef84a3b128aea2518b596)

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

gtk+: Explicitly disable xinerama since we don't have it as a DEPENDS

(From OE-Core rev: dbc25ca76d482f30186562fc51f5b3bdf48c0a7b)

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

gnome-desktop: Ensure we're deterministic about startup-notification dependency

Without this change we may or may not include startup-notification support.

We therefore explictly include it in the dependency list.

(From OE-Core rev: 8ad24306d8bc9c2fd73f4b814eb1a64c04707da5)

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

bitbake/lib/bb/msg.py: fix setting debug and verbosity levels

The debug and verbosity levels (as set by the -D and -v command line
options respectively)  were not being passed through within msg.py since
bitbake revision 45aad2f9647df14bcfa5e755b57e1ddab377939a due to
incorrect variable names.

Fixes [YOCTO #1513].

(Bitbake rev: c6e88b7c0e61f9586a275df53f48b90687c5f92f)

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

hob: store recipe path at load time

This fixes the internal dirtiness tracking such that if the Save menu item
is selected after loading a recipe the existing file is updated rather than
the user being prompted for the path to create a recipe at.

(Bitbake rev: 00fc1d7249b5e217cc7c36ac71b63ddad1c5b769)

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

hob: fix building with current selections after reparse

After the reparse we were setting the model to reflect the values before
the reparse was triggered but clearing the internal variables used to test
whether these values are set, leading to the UI erroneously reporting that
selections had not been made.

(Bitbake rev: 656eafe0f2c9ec7730d33e15705b8c720f787c49)

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

ui/crumbs/hobeventhandler: fix variable name typo

(Bitbake rev: f7d0560307707fe10bf80820f1e6ae300864f915)

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

ui/crumbs/hobeventhandler: move remaining getVariable calls to init

Instead of calling getVariable commands each time the BBPATH and BBFILES
entries need testing cache the results as a member variable at object
instantiation.

Fixes [YOCTO #1521]

(Bitbake rev: 109e1597671dfb7222672e268190aabc727960ca)

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

attr/acl: add SSTATEPOSTINSTFUNC

Added a native sstate post install function to fix the links
created between /lib and /usr/lib for the library files. These
links could point to an invalid build area when using shared state.

(From OE-Core rev: 8ab7b681cdb43c6c21c187b8cd01faa39727824a)

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

populate_sdk_rpm: add pkgconfig(pkg-config) to the list

(From OE-Core rev: 368b150416688654e35229a63b87177b52e83d02)

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

populate_sdk_rpm: add items to the INSTALL_PROVIDESNAME_RPM list

This fixes a problem when building meta-toolchain-gmae, by adding items that
will be provided by the host system, such as /bin/bash, /usr/bin/env and
libGL.so

(From OE-Core rev: 01361f9d25b0a0027bbbe713b93051a4663b14fc)

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

intltool: add libxml-parser-perl-native dependency to -native version

Fixes [YOCTO #1514]

Without a native dependency on libxml-parser-perl-native,
shared-mime-info-native can fail its do_configure task.

checking for XML::Parser... configure: error: XML::Parser perl module is
required for intltool

Testing: Successfully built shared-mime-info and shared-mime-info-native for
qemuppc.

(From OE-Core rev: 51b1df89828e677232e125181209b26d3c5ec928)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
CC: Joshua Lock <josh@linux.intel.com>
CC: Richard Purdie <richard.purdie@linuxfoundation.org>
CC: Koen Kooi <koen@dominion.thruhere.net>
CC: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

multilib.bbclass: Partially fix multlib image targets

This patch partially fixes problems when building multilib extended
images such as libXX-core-image-minimal. Its not a perfect/complete
solution but works much better than any previous code did.

[YOCTO #1496] (partial)
[YOCTO #1497] (partial)
[YOCTO #1498] (partial)

(From OE-Core rev: 00c38774ef0232cc2be924ed8e59220e7c452096)

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

qemu-config: use pkg_postinst to generate the proper shutdown.desktop

[YOCTO #1507]

We need to remove the file qemuarm/shutdown.desktop, or else, on qemuarm,
due to the PACKAGE_ARCH overriding from all to qemuarm in base.bbclass,
the generated deb file will be stored at
tmp/deploy/deb/qemuarm/qemu-config_1.0-r21_allarch.deb rather than
tmp/deploy/deb/all/qemu-config_1.0-r21_all.deb, and the package qemu-config
won't be installable -- task-base finally rdepends on qemu-config, so we get
the do_rootfs failure:

The following packages have unmet dependencies:
|   task-base-extended: Depends: task-base but it is not going to be installed
| E: Broken packages

There is also a generic shutdown.desktop, we can keep it and use a proper
pkg_postinst to cope with the case of qemuarm.

(From OE-Core rev: 751212d5effdceab91d95705e647cf07e6820940)

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bluez4: split gstreamer plugin into a seperate recipe

This eliminates a good chunk of buildtime when building console-only image.

The bluez4 and this recipe share a .inc file since the source is the same.

(From OE-Core rev: 5eaed4bba35e8188855acb93344adf48201748a2)

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

diffstat: Add missing file from previous commit

(From OE-Core rev: 6f4e6d6d41f874844b186b9e5b63a1b851becb52)

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

diffstat: Fix a build failure when using libdir=/usr/lib64

(From OE-Core rev: 9a846d83a39339de6d7cc0da050a50d7f4e093c7)

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

gnome-desktop: Fix python path in scripts to use target path

(From OE-Core rev: 22fd67f854f70f79ea94af11c61ef63939d54ac2)

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

foomatic-filters: Disable filter checking

Disable a few checks, and hard code the values for a few other items
to work around potential host contamination issues.  We also default
to the cups configuration for items.

(From OE-Core rev: ef7ad7a7c6b5f5a6c9d0d9e35e07b5a0dabe5be9)

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

cups: Fix recipe to use the correct cups directories

${libdir} is not used, instead they use a common ${exec_prefix}/lib
directory structure for helpers, filters, renderers, etc.

(From OE-Core rev: 24ae432b1a3906956381d83c1984687e45c5a1d1)

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

eglibc-2.12: Fix build on powerpc/603e

We pass --with-cpu to eglibc now. Which breaks
the configure for cpus that it does not support
We add support for ppc603e which gets 2.12
building for qemuppc.

(From OE-Core rev: 465a988e2370ec377875b599045f2a7bad913ac6)

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

Revert "eglibc: fix packaging of some of nativesdk packages"

This reverts commit 974cdb88680359aac5aefd11b84c4a374ab3a753.

See the mailing list for details, this broke things for too
many people such as where PKGSUFFIX is used in PREFERRED_PROVIDER.

odified:   meta/recipes-core/eglibc/eglibc-locale.inc

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

glib-2.0: Drop old 2.27.5 development version

(From OE-Core rev: 468e4a2f0e63e9d33e88b98a7d075b186ef481b2)

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

qemugl: Use local variable rather than "push" to save register

New gcc uses "%esp" rather than "%ebp" to index local variable in stack, and
push between save-to/restore-from stack decrease "%esp", which leads wrong
index. Saving registers via local variables to make gcc aware of this and avoid
stack disorder.

[YOCTO #1442] got fixed

(From OE-Core rev: afc9edc27e77e80fdd24b4c8c538f91672940e75)

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

eglibc: fix packaging of some of nativesdk packages

eglibc.inc contains packaging rules that depend on PKGSUFFIX expansion.
However that variable wasn't expanded properly in some case. E.g.

PACKAGES = " ... nscd${PKGSUFFIX} ..."
FILES_nscd${PKGSUFFIX} = "${sbindir}/nscd*"

would be expanded to

PACKAGES = " ... nscd-nativesdk ..."
FILES_nscd = "${sbindir}/nscd*"

Thus leading to empty -nativesdk packages and missing files. To overcome
that use python code blob to properly define PKGSUFFIX variable.

(From OE-Core rev: 974cdb88680359aac5aefd11b84c4a374ab3a753)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: include plugin-related headers into packaged SDK

Include headers necessary to compile gcc plugins into cross-canadian gcc
packages.

(From OE-Core rev: d12aa92b3dac1109d510e7b6f74055d1ab927817)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

e2fsprogs: include devtools scripts

Some scripts are necessary to develop programs with libcom_err and
libss. Include those into e2fsprogs-dev package.

(From OE-Core rev: 46332c2313abb273f6fc889fac4daa91cf43faa3)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glib-2.0: include glib-gettextize stuff

glib-utils already includes glib-gettextize program. Include some files
necessary for glib-gettextize to work.

(From OE-Core rev: c98356e9c46cd28b7ca8e84fe0ea56dc6d812a8d)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

alsa-lib: Don't use versioned symbols on uclibc builds as it causes strange hangs

alsa-lib: Don't use versioned symbols on uclibc builds as it causes strange hangs. Taken from oe-classic.

(From OE-Core rev: b354eb957ce08ac7814ce46c13ca3a8449b4063a)

Signed-off-by: Daniel Lazzari Jr <dlazzari@leapfrog.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libxml2: reinclude a -staticdev package

libxml2 provides a libxml2.a library, which wasn't included in any of
the packages. Usually static libs are included into a -staticdev
package, but this package wasn't generated for libxml2, due to
libxml2.inc redeclaring PACKAGES variable (to include libxml2-utils).
Replace PACKAGES setting with expansion to include all default packages
(and so -staticdev package with libxml2.a).

(From OE-Core rev: 76052861cc95fd4ad4c4b9eb6ce4cd1065ad4dc9)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gettext: shut up unshipped files warning

It seems nobody uses preloadable_libintl.so in OE world. It's not
included in any package. To stop oe-core from emiting a unshipped files
warning, remove that file in the end of do_install.

(From OE-Core rev: 9d33b6973438f514f1ca609c3a936b45af921e2f)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

polkit: fixup packaging

1) Drop static versions of extensions, as there is no point in them
2) For completeness include .la files for extensions into polkit-dev

(From OE-Core rev: e1a9f9e7f823330ab5140735b80988541e86e2d4)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gtk+: packaging fixups

1) Package gail modules into libgail package. .la files go into main
gtk+-dev package, debug symbols to gtk+-dbg.

2) For completeness add printerbackends .la files to gtk+-dev

(From OE-Core rev: a797c4ff1246566b545dc6cab2713a04f24fea6e)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libgpg-error: drop common-lisp files

OE doesn't have common-lisp neither in oe-core, nor in org.oe.dev. Stop
this package from installing clisp-related files.

(From OE-Core rev: 894254c18fe19b6321c790d9a5d93b24868ba6f2)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

avahi: include dbus interfaces definitions into daemon package

(From OE-Core rev: 8b7cd9a2e3a38fb5ed8207b2a6395855e1482ae3)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

insane.bbclass: don't warn on .so files in -nativesdk packages

There is no point in warning against .so symlinks in -nativesdk
packages. They have to contain such links, so shut up the warning.

(From OE-Core rev: 287d02a731e45ce5d843db1fd4d78ba17e9ed487)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

icecc.bbclass: fix handling of nativesdk and cross-canadian packages

icecc.bbclass assumed that nativesdk and cross-canadian packages should
use local icecc version. Instead those packages should use target-like
scheme for toolchain tarball generation.

(From OE-Core rev: 394c7c8cf262efeaafce40dc5eb480a1e041ac50)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

icecc-create-env: rename to -native

icecc-create-env(-native) is a native package. But if it doesn't have
the -native suffix, some dependency tracking stuff in oe-core (in
nativesdk.bbclass) get's crazy about it, trying to generate dependencies
to icecc-create-env-native packages. To fix that simply add -native
suffix to this package.

(From OE-Core rev: 83f9ebe9afb8bec4531899f8001a585c974e1db4)

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python: Fix cgi.py runtime issue

By default cgi.py attempts to use /usr/local/bin/python as its
interpreter.  However, on my Linux systems, including OE-Core,
python is installed into {bindir}.  Adjust this one file based on
the comment at the top of the upstream file.

This resolves an issue where a runtime dependency discovered during
RPM packaging breaks the rootfs construction.

(From OE-Core rev: 1da01664963b9c6a6df171dcd0fbb1406544035b)

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

task-core-lsb: Remove perl-dev

Remove the perl-dev recipe from the required packages.  When this
recipe is included, libc6-dev is automatically added as are a large
number of other -dev packages.  Eventually this creates a system
where the dependencies may not be able to be satisfied.

The perl-dev package should not be required in order to pass the LSB
testing.

(From OE-Core rev: 3f24b0f82458bcf15eabc74b9f370d9e774d0f63)

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

sudo: Fix PAM dependencies

When PAM is enabled for the distribution we need to be sure to have a
build dependency of libpam, and runtime dependencies on the pam modules
used by the sudo package.

(From OE-Core rev: 79db5ccc7ee9b360451074dd42f7676e138cf384)

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

perl: Fix a few perl binaries to use target interpreter paths

A small number of target installed programs, "${bindir}/pod*",
were incorrectly using the perl-native paths to access perl.  This
caused a dependency failure during RPM rootfs install, and also
prevented those specific components from functioning.

(From OE-Core rev: 7b8db36e088e3e9cd585e8d8dde1c1c2860c3c2e)

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

usbutils: Move the pkgconfig file(s) to the -dev package

pkgconfig files belong in the -dev package.  They automatically add
a runtime dependency (when packaging with RPM) for "pkgconfig", which
is not desired to be installed with this package.

(From OE-Core rev: 79cd09017693e0fd1634ea9748c797fcec4dac14)

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

bluez4: Replace deprecated configure switch

"--enable-configfiles" is deprecated upstream. Here is the original reason &
patch:

http://www.spinics.net/lists/linux-bluetooth/msg14266.html

Replace "--enable-configfiles" with "--enable-datafiles" in the bluez4
recipe.

(From OE-Core rev: 95e73d81554969ca4f0af43773343e1fbceda999)

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

libglade: fix the python environment settings

@PYTHON@ will points to the abstract path in sysroot, which should be
replaced by "#!/usr/bin/env python".

This should fix the sato-sdk rootfs error.

(From OE-Core rev: 51cd0bdf1fad230e919392db8c72afe2d9e21fdb)

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

glib-2.0: Fix python paths in scripts to use target paths

(From OE-Core rev: 08fd9740b9359425ec6ee9810fab1152e9737b0c)

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

procps 3.2.8: remove initscript

Sysctl is managed by other initscripts in the sysV world and by systemd itself in the systemd world, so drop it from the procps recipe.

This also fixes some style issues and a global FILES assignenment.

(From OE-Core rev: 80650bf4bce83e4e8d785e9a69eac0abbcdc6844)

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

xserver-xf86-lite.inc: use ${datadir} instead of /usr/share

(From OE-Core rev: d86acde7c6c57bd848feb1875c10df9b1de1201d)

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

xserver-xf86-dri-lite.inc: use ${datadir} instead of /usr/share

(From OE-Core rev: e83f142e1510cf686d9626b873ce9acf7baced0a)

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

xserver-xf86-common.inc: use ${datadir} instead of /usr/share

(From OE-Core rev: 9e283b637fcfc704156675eef4bc53966a03e3dc)

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

linux-yocto: kernel configuration cleanups

There were some config fragment warnings that were being hidden
by virtue of being output while generating the full config versus
the ones that were generated during the configuration audit task.

This change does the following:

  - Updates the kern-tools SRCREV to include changes that move
    all kernel configuration warnings to the audit script

  - Updates the meta SRCREV for linux-yocto to get updated
    kernel configuration fragments that are warning-free. These
    are duplicated configs, and other changes that have a net
    result of no runtime changes.

(From OE-Core rev: 8e67d4dd1e1bc64a84791abc613e237e29dd2aaa)

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

linux-yocto: add blktrace feature

Updating the 3.0 kernel to enable blktrace to match our previous
functionality. This is done by the following meta branch commit:

[
    meta: add blktrace feature

    Add a 'blktrace feature' that turns on kernel support for blktrace, a
    block I/O tracing tool.  Added to 'standard' alongside similar tracing
    tool config.

    (From OE-Core rev: 21654cbf766f9c319b2fc8a08f0060c60ef005eb)

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
]

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

linux-yocto: simplify kernel configuration check expression

kernel configuration validation takes place between two files. An
unprocessed configuration file (which is all the options found in
the various configuration fragments) and the final .config produced
by the lkc.

The unprocessed configuration file's name historically is based on
the name of the branch that was used to build the BSP. But with the
ability to map machine names to arbitrary branches, this is no longer
always true.

Searching for the pattern *-config-* in the meta subdirectory will
only match the config file, and frees the config validation phase
from being concerned with what branch was used to build the BSP.

(From OE-Core rev: 92e0d61f5bf15ca4eb262dfa3c533f9209a87915)

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

distro tracking: Update Distro Aliases

(From OE-Core rev: 40b52d9ed9b7b39bc86b3c87ae60fa6359cac265)

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

rootfs_rpm: two step installation

Firstly we install a base image.
Then install those multilib packages listed in MULTILIB_IMAGE_INSTALL.

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

Fix mercurial fetcher in fetch2

The _build_revision method in Hg class gets called with the wrong number
of arguments. This tiny patch adds a 5th argument to the method
declaration to prevent python from throwing an exception.

(Bitbake rev: 623e9c7f7a9cf12b8c81c26cc608990682a601dd)

Signed-off-by: Julian Pidancet <julian.pidancet@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

hob: fix opening of image output dir on image build completion

Firstly, rather than polling the DEPLOY_DIR_IMAGE directory each time an
image is built store the variable as a member of the hobeventhandler.
Secondly emit the generic "build-complete" signal *after* the specialised
"build-failed" or "build-succeeded" signals such that the appropriate
state variables are set before we try and use them.

(Bitbake rev: 043914a8b478fd4a7799acd1b44bdb3b0af2165a)

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

ui/crumbs/runningbuild: fix log messages right-click menu

Pass the correct callback name to the connect method.

(Bitbake rev: d2b9167a7b0eb3fcb31d64f915beef41aac659a5)

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

libpng: Add RPROVIDES for backwards compatibiliy with previous libpng packaging

(From OE-Core rev: eb39d55d07880e9338808fc077991b5673077aa1)

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

debian.bbclass: Fix hardcoded "lib" reference

This patch removes a hardcoded "lib" reference from debian.bbclass
allowing libdirs like "lib64" to be used. Whilst working on this area of
code, a similar assumption about "bin" is removed and the regular
expression compilation is moved to the parent function since it doesn't
change and this will improve the speed of the code.

(From OE-Core rev: 6b2e41023941f15d0f9211ef17339ba1fca9a3b9)

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

Update python dependencies to be simply to "python"

Previously python dependencies were of the format "python(abi) = ..."
This format is not yet supportable within OE, so revert to a form
we know we can handle.

Introduce a change to package.bbclass that ensures it will causes the
sstate-cache's "package" to invalidate.  Since pythondeps changed, the
output of rpmdeps changes, which causes the per-file dependency
information to change.... thus we need to invalidate the cache!

(From OE-Core rev: 30064a98dc9049db4a37f119d15fbb59aa3c8377)

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

libpng: avoid dangling link in libpng package

As for libpng recipe, the libpng.pc is a symbol link to libpng12.pc,
if packaging them into two different packages, we will get a dangling
link for libpng.pc, and when calculating rpm dependency with that file,
we will get nothing.

Only keep one version of dev package to fix the issue.

(From OE-Core rev: 50232671951e0b69d52a685b82a10be335458292)

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

oprofileui: Change avahi to avahi-ui depends

(From OE-Core rev: c89aa6a2b22e05e98991cc5b56d6ec869cf9c69a)

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

Fix alsa-utils package dependencies

Make alsa-utils main package depend on all of the more specific alsa-utils packages (such as alsa-utils-alsamixer).

(From OE-Core rev: 181063d0ae52b6c5ec203ea7c4fee629761cafdf)

Signed-off-by: Daniel Lazzari Jr <dlazzari@leapfrog.com>
PR Bump

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

multilib_global.bbclass: Fix non-multilib package provides

In non-multilib packages, configured in a multilib configuration we
need to adjust the system provides and rprovides to include the
virtual multilib variant.

This resolves a problem introduced in the
329d864f9bbf94ad3aae8df43d63fe10e4237e4f commit.  Where "allarch"
packages were suddenly providing all variants of an object.

(From OE-Core rev: 66fa6b7e13fbcc5f75fb1b8aa3aedfbdbc148688)

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

busybox: Enhance to add dynamic per-file provides

When using the RPM package backend, we need a full list of per-file
provides (and requires).  The busybox package provides a number of
command line utilities, such as /usr/bin/env.  However, because
the utilities are created at post install time via scripting the
provide of each of the links was never made.  So any programs,
such as python, that require /usr/bin/env were unable to resolve
the dependency and failed.

This change only affects packaging backends that use per-file
dependency data.  Currently RPM is the only packaging backend
with this ability.

(From OE-Core rev: dd63f64155e0fe41c27a8c9be84dfc0a255a3ff7)

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

Add a run-time dependency that eglibc support GNU_HASH

RPM checks to see if binaries require GNU_HASH, if they do it adds
an automatic dependency of "rtld(GNU_HASH)".  We need to satisfy
this dependency, and we do it by providing rtld(GNU_HASH) in the
package(s) that provide a runtime-linker.

(From OE-Core rev: e150afea5a9af8e1bc5aaecd9eb3fe86d7b79f44)

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

Fix RPM dependencies

When packaging for RPM, dependencies were not being properly added
to the packages.  Only the "providing" dependencies were added due
to a typo.

(From OE-Core rev: c5d4e17f29d7f032d566d3b222a0ee231da03f78)

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

multilib: install MULTILIB_IMAGE_INSTALL

If user set MULTILIB_IMAGE_INSTALL, we need to install those multitlib
packages into the final image.

Also fix the logic in handling multilib prefix. For certain case like a
normal image contains several multilib libraries, the image recipe isn't
extended with MLPREFIX, therefore we need to enumerate the possible
multilib prefixes and compare them with package prefixes.

(From OE-Core rev: a94cc552dc764d3cc83ea638cb08b9a2a8829d2f)

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

rpm: add multilib prefix for archs under deploy/rpm

Currently MACHINE_ARCH deploy folder is unique in multilib system, thus
a lib32 version of rpm package will override a normal rpm package if
its PACKAGE_ARCH is ${MACHINE_ARCH}.

Define different deploy folder for multilib architectures to avoid the
confliction.

(From OE-Core rev: 921f984aa65e23d5a8ec5c2e58a96cb8a4790b5d)

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

avahi: Refactor recipe to move UI to avahi-ui

[YOCTO #1492]

This patch refactors avahi into a core non ui and avahi-ui, OE-Classic
has a similar factoring, and it was used as a refernce to make this
occur correctly here.

This allows the package to be built without gtk and GPLv3 issues.

(From OE-Core rev: a3909a6a42a1e3e80b7f2597f95ecb7db2814d2d)

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

distrodata: fix distro_check code

This fixes a problem with package names from inherits showing up

(From OE-Core rev: edb2a11994ac21e790155ea519bded4b37ef9307)

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

zypper: git repo moved to github

(From OE-Core rev: 924d2ce0326b17a18909ec9eea8b5f75e497301a)

Signed-off-by: Anders Darander <anders@chargestorm.se>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sat-solver: git repo moved to github

(From OE-Core rev: cdaa7e00fb11f3b583d765eb3d578404998e18a9)

Signed-off-by: Anders Darander <anders@chargestorm.se>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libzypp: git repo moved to github

(From OE-Core rev: 7e67aa4002c50cd91e39e5e9305f8cb86c8c0f11)

Signed-off-by: Anders Darander <anders@chargestorm.se>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

deb packages support: switch from /var/dpkg to /var/lib/dpkg

[YOCTO #1086]

The pach was backported from OE:
http://git.openembedded.net/cgit.cgi/openembedded/commit/?id=41e0fbf792037f249d1b8d283b3de81718887c9f

(From OE-Core rev: d1e6f49a6473df3c626100ba01b27485f735c33b)

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scripts/combo-layer: fix still overzealous regex in default hook script

In the previous fix to this hook script (OE core revision
e7aae45414e4597e9244f86a81fbc940f73785c8) a start-of-line (^) marker was
missed, so if a commit had no Signed-off-by line but it contained an
inner patch that did, the inner patch was modified causing a "corrupt
patch" error.

(From OE-Core rev: 9f1b4e5923c60cd4215b0ca4da2cab6245e54ccb)

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

hob: enable package only builds even if an image has been built

Once an image build has been performed the selected_image variable would
prevent the package only build choice being triggered - by setting the
variable to None on reset we are able to perform package only builds after
an image build.

(Bitbake rev: 705156e2812afb288632f01416fcbbf9add26bee)

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