acl: fix the order of expected output of getfacl

The result of getfacl is sorted by user id.
In Centos or RHEL, bin user id is 1 and daemon user id is 2.
But in our image, bin user id is 2 and daemon user id is 1.
The patch fixes this issue to make ptest pass.

(From OE-Core rev: a5180e942c9315d280580773e72fe67f27629a3c)

Signed-off-by: Chong Lu <Chong.Lu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: output filenames based on initial recipe name

If a recipe for an image was edited from the hob interface the name
of the files outputed in the <build_dir>/tmp/deploy/images/${MACHINE}/ and the
temporary recipes from <build_dir>/recipes/images/ contained only the
generic name "hob-image". From now on both the temporary recipes and
the output from the deploy/ directory will contain the name of the base
recipe appended by the "-edited" suffix, in the case when a base image recipe was
edited. The base recipe can be a standard recipe (e.g core-image-minimal) or
a custom created and saved by the user.

For example, if core-image-minimal is edited the deploy/ directory will contain
core-image-minimal-edited-20140318-140428-qemux86.ext3 and the recipes/images/
directory will contain the recipe core-image-minimal-edited-20140318-140428.bb.

[YOCTO #5002]

(Bitbake rev: f34575809677dc52e1071a3ae3daebe92819cec0)

Signed-off-by: Marius Avram <marius.avram@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: fetch2.URI: Set username/password should not change the other

When setting the username after already having set the password, the password
was unexpectedly reset. This change fixes this issue and introduces unit tests
to make sure it doesn't happen again.

(Bitbake rev: 25faef3a047f9c7564089463d7c96f6910b640cb)

Signed-off-by: Olof Johansson <olof.johansson@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: fetch2.URI: Support URIs with both query strings and params

There is a case in meta-intel where a SRC_URI contains both a query string and
URI parameter:

 https://edc.intel.com/Download.aspx?id=6190;downloadfilename=LIN_IEMGD_1_14_GOLD_2443.tgz

Python's urlparse thought the URI parameters were part of the query parameter
value, but in the bitbake context this is obviously not the case. As bitbake's
usage isn't really RFC compliant, we have to extract and remove the URI parameters
*before* urlparse sees the URI.

(Bitbake rev: c2f27ae4271985b48f957747b6ea372c8699cb49)

Signed-off-by: Olof Johansson <olof.johansson@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: fetch2.URI: add support for query parameters

This change introduces the .query property of the URI class. It is a
read/write dict of the parameters supplied in the query string of the
URI. E.g.:

  http://example.com/?foo=bar => .query = {'foo': 'bar'}

(Bitbake rev: 1cb2b3c458c8c5521591d2c8f2e0058143fc77bb)

Signed-off-by: Olof Johansson <olof.johansson@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: tests.fetch: Remove debug assert

(Bitbake rev: f112660bca0ed8be061055b1e388deeb2d1980a7)

Signed-off-by: Olof Johansson <olof.johansson@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: fetch2.URI: Coerce urlparse to use netloc for all schemes

(Bitbake rev: 4d502578f022bcf772780550c047b8c09ba01443)

Signed-off-by: Olof Johansson <olof.johansson@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

busybox: disable fsck.minix and mkfs.minix

The MINIX and MINIX 2 filesystems are not really used anymore.

(From OE-Core rev: 4a47c943a76e79af82d47d8fa83c6687e0d915a0)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/image: ignore modules.* changing during multilib image construction

Since we now run depmod when building images (as the postinst that does
this is now on kernel-base instead of kernel-image) it is possible to
have module file differences between the two halves of the multilib image,
and the code that checks for such differences detects this and fails.
Whitelist this file to avoid the failure.

Specifically, modules.alias, modules.dep and modules.symbol can differ
along with their .bin counterparts.

Related to fix for [YOCTO #5392].

(From OE-Core rev: 0a315804bf991664c0948e3024b8e8b9e9085808)

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

classes/kernel: move module postinst commands to kernel-base

Since kernel-base is the package that contains the files that depmod
needs to run, we should be running depmod from the kernel-base
postinstall rather than kernel-image.

Fixes [YOCTO #5392].

(From OE-Core rev: f7d2cb383281ec8dfa90950ba04d87dd29ffc676)

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

gtk-engines: removed gnome inheritance

Remove gconf class (via gnome) to fix WARN-QA message regarding
--disable-install-schemas option.  Add missing intltool-native build dependency
that previously came through gnome.bbclass.

(From OE-Core rev: ef2ba0a17a4fbd0e0c43c144f43d1d40de23697e)

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

classes/image: disable pam nullok_secure with debug-tweaks

If you want passwordless logins to work with pam enabled, then you can't
have "nullok_secure" enabled on pam_unix entries. Add some
postprocessing to change these to "nullok" when debug-tweaks is in
IMAGE_FEATURES, in order to make passwordless logins with PAM work
again.

Fixes [YOCTO #5973].

(From OE-Core rev: a973fe10305c7d080d5c017ddd9b2bd860221659)

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

archiver.bbclass: move a few code to copyleft_compliance.bbclass

Move the code which is only used by copyleft_compliance.bbclass from
archiver.bbclassc, and remove the "inherit archiver" from
copyleft_compliance.bbclass.

The archiver.bbclass is used for archiving various types of sources, but
the copyleft_compliance.bbclass is used for analysing the license, they
don't have much relationships.

[YOCTO #4986]
[YOCTO #5113]

(From OE-Core rev: 578830fe2ff279ea620916ea711b80dc1b29a275)

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

package_rpm.bbclass: archive the source to srpm package

The archiver.bbclass will put the sources to ARCHIVER_OUTDIR according
to configuration, then the rpmbuild -bs will create the srpm.

[YOCTO #4986]
[YOCTO #5113]

(From OE-Core rev: f9ba047afb8780c8bd7cb1ba45470d30abf92e92)

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

archiver.bbclass: refactor it

The archiver didn't work, and there were a few problems, for example:
1) There was no src_dir.org (or orig), but the diff command still use it
2) There were a few duplicated code
3) It didn't archive the source for the native or gcc
4) The work flow is not very well
5) The "subprocess.call('fakeroot cp xxxx'" should be removed
6) And others ...

So that we have to refactor it, the benefits are:
1) Fix the problems and make it work well.
2) Reduce more than 300 lines
3) Make it easy to use.

Hre are the explanation about the bbclass:
This bbclass is used for creating archive for:
1) original (or unpacked) source: ARCHIVER_MODE[src] = "original"
2) patched source: ARCHIVER_MODE[src] = "patched" (default)
3) configured source: ARCHIVER_MODE[src] = "configured"
4) The patches between do_unpack and do_patch:
   ARCHIVER_MODE[diff] = "1"
   And you can set the one that you'd like to exclude from the diff:
   ARCHIVER_MODE[diff-exclude] ?= ".pc autom4te.cache patches"
5) The environment data, similar to 'bitbake -e recipe':
   ARCHIVER_MODE[dumpdata] = "1"
6) The recipe (.bb and .inc): ARCHIVER_MODE[recipe] = "1"

All of the above can be packed into a .src.rpm package: (when PACKAGES
!= "")
ARCHIVER_MODE[srpm] = "1"

[YOCTO #4986]
[YOCTO #5113]

(From OE-Core rev: e5b20199ddddfa2146dc112c0a81f3fd292258e5)

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

classes/archive*.bbclass: remove archive-*-source.bbclass

Nearly all of the codes in the following 3 files are the same, we can
move the code to archiver.bbclass and remove them:

  archive-configured-source.bbclass
  archive-original-source.bbclass
  archive-patched-source.bbclass

[YOCTO #5113]

(From OE-Core rev: d2d3428a0fc746e56c0a0046bd12a393dedb4dda)

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

bitbake: toaster: fix task elapsed time calculation

This patch restricts the elapsed calculation to just
the events that have the "time" parameter set.

This fixes an error where data was lost due to an exception
where invalid dictionary lookups were made on the wrong
events.

(Bitbake rev: fa9f4eb8784553deb782bff34c5e04012c2c52c9)

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

bitbake: toaster: image information views

[YOCTO # 4346]

When a target image is selected, this commit adds to the toaster
project a two-tabbed page that shows
1) 'packages included' a table of packages included in the image
(see target.html), and
2) 'directory structure', the target image's file system directory
and detailed information showing the source of each file in the
directory table (see dirinfo.html).

The directory structure tab relies on the open source jQuery plugin
jtreetable which provides hierarchical table expansions and contractions
of the directory entry tables as the user drills down into directories.

A file of jtreetable styles that are compatible with other toaster styles
is provided included as css/jquery.treetable.theme.toaster.css. The
complete unaltered jtreetable plugin is added via a separate commit.

This work was developed base on the bugzilla specification number 4346
and the document "Design 1.1 Image information" attached to that report.

Whitespace and typo fixes from Alex Damian.

(Bitbake rev: 1ba9f310a8b4fd0952a95be86ab43ae27fe6d983)

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

bitbake: toaster: add jquery treetable plugin

This commit includes the critical javascript and css files
downloaded from http://plugins.jquery.com/treetable version 3.1.0.
The following is from the author's README.md included with
the license files in the directory static/jquery-treetable-license.

"jQuery treetable is a plugin for jQuery, the 'Write Less, Do More,
JavaScript Library'. With this plugin you can display a tree in an
HTML table, e.g. a directory structure or a nested list...
Download the latest release from the jQuery Plugin Registry or grab
the source code from Github. Please report issues through Github
issues. This plugin is released under both the MIT and the GPLv2
license by Ludo van den Boom...
See index.html for technical documentation and examples. The most
recent version of this document is also available online at
http://ludo.cubicphuse.nl/jquery-treetable. An AJAX enabled example
built with Ruby on Rails can be found at
https://github.com/ludo/jquery-treetable-ajax-example."

Author: Ludo van den Boom

The relevant files were moved to the static/js and static/css
directories to match the existing javascript and css toaster
deployment model.

The applicable licenses and the README are in
static/jquery-treetable-license.

(Bitbake rev: 9a260d8b08053e3dcc2fe5960e060b2da1109790)

Signed-off-by: Dave Lerner <dave.lerner@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: filter out setscene tasks in recipe details

Filter out any *_setscene tasks in recipe details page, as
they are not relevant.

[YOCTO #5913]

(Bitbake rev: 3aca83accd6755d518d6b6c667e685020efdfb8d)

Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: Amend failed tasks behaviour

In the builds table, when a build reports more than one
failed task, the number of failed tasks should be a link
to the tasks table with the failed tasks filter applied.
Amend the URL to include the filter.

(Bitbake rev: 7d93dd7a1ddb43279e0e1309ac6d935bf6b2ec97)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: Set display of "Cache attempt" column

The "Cache attempt" column should display by default in the
tasks table, but should be hidden by default in the time,
CPU and Disk I/O tables.

(Bitbake rev: dc96eecd2e8bef1b4d4fc500bf496b3edb91f808)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: Remove extra space in the breadcrumb

For builds with more than one target, remove the extra
space between the '+' and the number of targets - 1.

(Bitbake rev: 0d5e342d8878d8aff0ffe723be376845cccd8720)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: Change "Package version" to "Version"

We use "Version" everywhere else.

(Bitbake rev: 9def611e25c0d90285012cf44cb40256984368f3)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: Remove trailing spaces from 'name'

Those extra spaces were showing in the heading of the
filter modal dialogs.

(Bitbake rev: ed958d288de92faf8bd17a067c7a7f719dacd27d)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: Add link to the Toaster manual

The Toaster manual will live at
http://www.yoctoproject.org/documentation/toaster-manual
and should open in a new tab / window.

(Bitbake rev: 4e382ac6b5c1da497caa1ece3fcdc404dd02c058)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: measure task duration with server-side timestamps

The buildstats and toaster use separate time markers to measure the
duration of task execution. This causes a mismatch in the time
measured by buildstats class and the time measured in toaster.

The solution implemented here is to timestamp the creation of
every TaskBase event on the bitbake server side and calculate
the execution duration as the difference between creation time
of TaskSucceeded and TaskStarted events.

Based on an original patch by Marius Avram.

[YOCTO #5485]

(Bitbake rev: 7a08282c074c264f414cf7665dc873f51245072c)

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

bitbake: ConfHandlerr: Use full path in variable history

When an evaluation was made for a configuration file the path to the
file was saved as a relative one. The change in this commit will save the
location as an absolute path. This way the user will have full information
regarding the location of the file where a variable was changed and the
line withing the file.

[YOCTO #5562]

(Bitbake rev: df9e22901555b06fef308f7136547f2c47ccec35)

Signed-off-by: Marius Avram <marius.avram@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnome-doc-utils: removed gconf inheritance

In order to remove WARN-QA message regarding --disable-schemas-install configure option.

(From OE-Core rev: eb53ae4a5702dbd65b6e340bbd3ae4566157c081)

Signed-off-by: Alexandru Palalau <alexandrux.palalau@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gsettings-desktop-schemas: removed unneeded class inheritance

Removed gconf, gtk-icon-cache and mime classes as they are not needed

(From OE-Core rev: 1138976f7bdcd24ef57451a2c2ceffa43666610f)

Signed-off-by: Alexandru Palalau <alexandrux.palalau@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnome-mime-data: removed gconf and gtk-icon-cache inheritance

Should remove WARN-QA message regarding --disable-schemas-install.
gtk-icon-cache is not needed.

(From OE-Core rev: 665d2190b21d43081db806a49bfd529544ee87f2)

Signed-off-by: Alexandru Palalau <alexandrux.palalau@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gst-plugins-gl: removed gconf inheritance

In order to remove WARN-QA message regarding --disable-schemas-install configure option.

(From OE-Core rev: 1bb8c4018c4c52db0934ffcd7cfdf5d5b599a9f1)

Signed-off-by: Alexandru Palalau <alexandrux.palalau@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnome-desktop: removed gconf inheritance

In order to remove WARN-QA message regarding --disable-schemas-install configure option.

(From OE-Core rev: 0c30f3106a5172b1e58d5cc4eae964923fa6c8e2)

Signed-off-by: Alexandru Palalau <alexandrux.palalau@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

core-image-minimal: allow user to set size

Allow the user, in their local configuration, to override the size of the
final image. This is useful when creating an image for (for example) an SD
card and the user wants the image to fill up the card as much as possible.

(From OE-Core rev: 9b4ea0f2e13e52d860d59d0348a3218af151666b)

Signed-off-by: Trevor Woerner <trevor.woerner@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python3: sync module dependencies from 2.7

These have been added recently to 2.7 but were missing in the 3.3
script/inc file.

(From OE-Core rev: 4669afac1004a89e6b87ec46136ca3e7448700d4)

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

generate-manifest-3.3.py: sync descriptions with 2.7 version

(From OE-Core rev: 081bc11c347d11d285f2948127bca81a285ada84)

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

libsdl: add missing libglu dependency for openGL

Configure checks for glu.h to determine if openGL is available.

(From OE-Core rev: a7641a2bc3f3f7e661f71b17f91382fb9a1cde55)

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

kernel-arch.bbclass: add arm64 support to U-Boot architecture map

to define UBOOT_ARCH, we map kernel architectures to U-Boot architectures.
In the case of arm64 kernel, we should map to arm U-boot architecture.

This patch add the exception rule to the map_uboot_arch function.

(From OE-Core rev: 0b891265716c414ade29d587fc1a3c4ea7beadbe)

Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

runqemu: Add option for custom BIOS directory

Add support to specify a directory for custom BIOS, VGA BIOS and
keymaps as supported by qemu (-L option). Even though this can be
done through qemuparams, having this option provides better user
experience by not having to specify a long and cluttered path along
with other qemuparams that the user might want to specify.

This new options assumes first that the path provided is relative to
OECORE_NATIVE_SYSROOT and will check whether it exists before proceeding.
If not, it will treat the provided path as absolute. This provides
the user flexibility to use BIOS binaries generated inside or outside
the OE build environment.

(From OE-Core rev: d302f5683dd736ac4cd4b601a046d22000d41e68)

Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base-files: use /dev/root in /etc/fstab for systemd support

systemd does not recognize "rootfs" in /etc/fstab so the root
filesystem is not checked. As a result, the following message
is logged by journalctl:
systemd-fstab-generator[68]: Checking was requested for "rootfs", but it
is not a device

Changing "rootfs" to "/dev/root" in /etc/fstab allows systemd to
check the root filesystem when the kernel is booted with the root
filesystem mounted read-only.

[YOCTO #5950]

(From OE-Core rev: c509f948d9c575c45af8c5ed1cb1692c0ca5dade)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

tclibc-uclibc: Changing assignment to conditional variable assignment

Setting of a variable PREFERRED_PROVIDER_virtual/libc only if it doesn't have a value

(From OE-Core rev: d0e9f74e3f1322b58b78a9bc82f299d6b9da036f)

Signed-off-by: Andrey Belous <abelous@broadcom.com>
Signed-off-by: Nicolas Dechesne <nicolas.dechesne@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio: Upgrade 4.0 -> 5.0

bluetooth-proximity has been removed in 5.0

pulseaudio now differentiates between bluez4 and bluez5

(From OE-Core rev: c53d3a4f1f2d42437dc2985fe109039c843aa4dc)

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

sstate: list missing files for toaster

Toaster needs to record the attempts to restore
setscene tasks that don't have a sstate file.

We build a list of tasks for which we can't find an
sstate file, and if we're running under Toaster data
collection, we send it off with a MetadataEvent.

(From OE-Core rev: 109ae6c5c981610ab0d63d2c83dcd50b2e93276b)

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

toaster.bbclass: read the data needed for license manifest path

The license.manifest file is located in DEPLOY_DIR_IMAGE/licenses/
IMAGE_NAME dir. The data needed is collected after rootfs task.

[YOCTO #5649]
(From OE-Core rev: ff52c5ba15433f2b1e9723bf845e39da918ad59b)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-runtime: Build libatomic

GCC 4.8 includes a new runtime library, libatomic, which supports
atomic operations not supported by hardware or the OS. Build it,
so other packages can link against it, if needed.

(From OE-Core rev: a4dd6dfccee0be50d3addce3dd1bf903e051ad5a)

Signed-off-by: Cosmin Paraschiv <cosmin.paraschiv@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image_types.bbclass: use 4096 instead of 8192 bytes-per-inode

The image not correctly created if 'ptest-pkgs' is in IMAGE_FEATURES,
this is because there is no free inode left. We can use 4096 instead of
8192 bytes-per-inode to fix the problem, and most of the distributions
us 4096, such as Ubuntu, Suse, Fedora and CentOS.

There are another problems:
* There are error message when there is no free inode left if we run the
  mke2fs command manually, but they are not in log.do_rootfs.

* The image generation doesn't stop when error happens because mke2fs
  doesn't return failed for this case.

Will fix them in other threads.

[YOCTO #5957]

(From OE-Core rev: 09ab3a00598d06e3a1bf871811c2ac37359c74da)

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

buildhistory.bbclass: create proper dependency files for SDK

The old functions were calling the list_installed_packages() wrapper
function that only listed the packages in an image rootfs. Even for
target/host SDK. Also, a python crash was possible if 'bitbake -c
populate_sdk core-image-*' was called without calling 'bitbake
core-image-*' first. That's because the wrapper was always looking into
the image rootfs...

This commit fixes the problem and calls the right wrapper for image/sdk.

(From OE-Core rev: c1b1a6eb448aa1548e2ec669a9304b5a25bd8ba5)

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

image.bbclass, license.bbclass: adjust the name of list_installed_packages()

The old wrapper got renamed to image_list_installed_packages().

(From OE-Core rev: 118a2a44bbe5ed2e9bbd0012970686be454e5d4c)

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

rootfs.py, sdk.py: adjust/create the wrappers for creating installed packages list

Since we created a new PkgsList object that will deal with listing the
installed packages in a rootfs, use the new class both for images and
SDKs in the wrapper functions.

The old list_installed_packages() wrapper listed only the packages inside
an image rootfs. It didn't deal with target/host SDK rootfs's.

(From OE-Core rev: 8fc18e67504db5b6df3fdd239c6187a71af52656)

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

package_manager.py: create separate class for installed packages listing

This commit creates a new class that has the only purpose to generate
various listings of installed packages in the rootfs.

Basically, the methods involved in listing the installed packages, that
were part of each backend PM class implementation, were moved to this
new class.

This change avoids instantiating a new PM object just to get the list of
installed packages in a certain rootfs.

(From OE-Core rev: a7290ed13378826723d1edc7e828eab848eaad10)

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

linux-dummy: provide kernel-modules (empty) package

Like it is done in kernel.bbclass

(From OE-Core rev: 5df4e63e747028c3ce89f1f9ae01a766ae34dc3d)

Signed-off-by: Matthieu Crapet <Matthieu.Crapet@ingenico.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cogl-1.0 : Update to 1.16.2 release

(From OE-Core rev: 3eb2435a7c7c8108cb73945d5003a9fc7455befd)

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

gsettings-desktop-schemas: Updated to 3.10.1.bb

(From OE-Core rev: 0bbea90252b93d7e36527dd0db63597a3d4c0aa1)

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

cairo: Add 'egl' and 'glesv2' PACKAGECONFIG options

(From OE-Core rev: f405f6983b4f36db4ec1b89c30b1082aec2545f8)

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

scripts/cleanup-workdir: Fix a comment typo

 s/Romve/Remove/

(From OE-Core rev: 5f8924a8f76f314e3741dad6ab567452ae117114)

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

e2fsprogs: return error when failed to populate fs

We need return retval when "mke2fs -d" failed, otherwise the "$?" would
be 0 which is misleading.

[YOCTO #6011]

(From OE-Core rev: 46896b601c1c93e276954b674aa30b8b4dc4f611)

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

alsa-tools: fix build when x11 and gtk+ not available

 Updated patch to work with alsa-tools version 1.0.27

(From OE-Core rev: baada8ef412478450fa0ab234b1a3151bdfe7ac4)

Signed-off-by: Simone Agresta <simone.agresta@bticino.it>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

perf: add libexec and traceevent to packages

Recent versions of perf may install files into /usr/libexec/perf-core and in
/usr/lib/traceevent. To avoid packaging QA errors, we add these two
directories to the FILES variables.

We also add: INHIBIT_PACKAGE_DEBUG_SPLIT="1" to avoid the following issue
(due to a trailing / being removed):

   ERROR: debugedit failed with exit code 256
   ...
   debugedit: canonicalization unexpectedly shrank by one character

And finally, we must ensure that the traceevent libraries are installed to
the proper multilib library path. If building some multlibs, the incorrect
library path will be selected by perf, since it triggers via: ifeq
($(ARCH),x86_64) (or similiar mechanism per arch).

In a 32 bit build, with a 64 bit multilib, the arch won't match and the
detection of a 64 bit build (and library) are not exected. To ensure that
libraries are installed to the correct location, we can make the substitution
in the config/Makefile. For non multilib builds, this has no impact.

(From OE-Core rev: c1b5a262c0201faf2c6bf545d6acb32dfe383ba3)

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

fontcache.bbclass: add fontconfig-utils runtime dependency

In an image recipe, you can get a warning, for example:
WARNING: The postinstall intercept hook 'update_font_cache' failed (exit code: 1)! See log for details!
WARNING: The postinstalls for the following packages will be postponed for first boot: ttf-dejavu-sans-mono

(because /usr/bin/fc-cache is missing)

In OE-core, rdepend is correctly done in each recipe:
- ttf-fonts/liberation-fonts
- ttf-fonts/ttf-bitstream-vera

In meta-OE, rdepend is not done but packagegroup-fonts-truetype.bb includes fontconfig-utils:
- ttf-fonts/ttf-arphic-uming
- ttf-fonts/ttf-dejavu
- ttf-fonts/ttf-droid
- ttf-fonts/ttf-gentium
- ttf-fonts/ttf-hunkyfonts
- ttf-fonts/ttf-inconsolata
- ttf-fonts/ttf-liberation
- ttf-fonts/ttf-mplus
- ttf-fonts/ttf-sazanami
- ttf-fonts/ttf-ubuntu-font-family
- ttf-fonts/ttf-wqy-zenhei

(From OE-Core rev: 306335b13fedc18cf03da1c2a68e97c01eb59075)

Signed-off-by: Matthieu Crapet <Matthieu.Crapet@ingenico.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

genext2fs: remove it

Remove genext2fs since we don't use it anymore, it can't support
ext4 well, either. We have used "mke2fs -d" to instead of it.

[YOCTO #6013]

(From OE-Core rev: ff5666bc460520aef6105e117d5431c05fd9f55b)

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

generate-manifest-2.7.py: sync with python-2.7-manifest.inc

Some changes were made earlier by me to SUMMARY values in
python-2.7-manifest.inc without changing the manifest script.

(From OE-Core rev: 45779941cec4f53a8ca7f8350402e5d9e866c916)

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

generate-manifest-*.py: set SUMMARY instead of DESCRIPTION

Typically what's being set here is a short description, so it makes much
more sense to set SUMMARY.

(From OE-Core rev: d3941c88e2639637a8bc0b2c31c1d892d2ae40e3)

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

python: add python-mmap to python-multiprocessing RDEPENDS

The multiprocessing module needs the mmap module. We got away with not
having this for the build appliance because python-mmap was added to
packagegroup-self-hosted, but this is the proper place to have it.

(From OE-Core rev: 1c031c627a362b3f18ffa2e9caeb6cfb299b9948)

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

x264: add perlnative to inherit

perl is run as part of the configure process, so we should ensure that
it's available.

Should fix [YOCTO #5768].

(From OE-Core rev: ff7071418217c343276c22c7e1a053672f92639b)

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

useradd{-static}: Ignore useradds on nativesdk

The code was supposed to ignore both native and nativesdk operations when
using the useradd and useradd-static code.  However, somewhere along the way
the code was dropped.  This didn't cause any issues until someone enabled the
enforcing mode in the new useradd-static and various nativesdk packages
started to fail.

(From OE-Core rev: 6b9705892400a1da1fcd973c64d1911c7c4463f6)

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

bitbake: runqueue: Remove use of waitpid on worker processes

Use of waitpid on the worker processes is a bad idea since it conflicts
directly with subprocess internals. Instead use the poll() method
and returncode to determine if the process has exitted, if it has,
we can shut down the system.

This should resolve the hangs once and for all, famous last words.

(Bitbake rev: 60969cd62e21e7d4af161bf8504b7643a879c73f)

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

alsa-tools: fix build when x11 and gtk+ not available

 Updated patch to work with alsa-tools version 1.0.27

(From OE-Core rev: ef196434620522affc11b5b1b867386b5d14a4c3)

Signed-off-by: Simone Agresta <simone.agresta@bticino.it>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sgml-common-native: fix the generation of sgml-docbook.cat

There would be build errors if multiple builds use the same SSTATE_DIR,
and the builds use the different versions' docbook-sgml-dtd-native, the
problems are: we have multiple versions docbook-sgml-dtd-native:

docbook-sgml-dtd-3.1-native
docbook-sgml-dtd-4.1-native
docbook-sgml-dtd-4.5-native

And they depend on sgml-common-native which installs the file
sysroot/etc/sgml/sgml-docbook.cat, this file is also included in the
sstate cache file (.tar.gz), but both the 3 versions
docbook-sgml-dtd-native may update the
sysroot/ect/sgml/sgml-docbook.cat, and it is a hardlink to
SYSROOT_DESTDIR/etc/sgml/sgml-docbook.cat, so the sgml-docbook.cat in
the sstate cache file may contain the
docbook-sgml-dtd-<version>-native's info, and there would be errors when
we mirror the sstate-cache to another build which uses a different
version.

Now we exclude the sgml-docbook.cat from the ${D}, and generate
sysroot/ect/sgml/sgml-docbook.cat dynamically will fix the problem, both
the sgml-common-native and docbook-sgml-dtd-<version>-native can update
it it correctly.

The similar to docbook-dsssl-stylesheets-native and openjade-native.

[YOCTO #5994]

(From OE-Core rev: 50683df81e1605ad7c03bc633aa55da7e97cfe62)

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

expect: fix do_install failure on SLED 11.2

Previously we were encountering do_install failures on older machines
like SLED 11.2. These machines have relatively lower versions of glibc.

     tclsh8.6: /lib64/libc.so.6: version `GLIBC_2.14' not found

The above failure is due to that fact that we are using the fixline1
tcl script to install commands under the example directory. The purpose
of fixline1 is to ensure that the installed scripts begin with '#!/bin/sh'.
However, the scripts under the example directory have already got
the correct first line. That's why we don't need to modify anything and could
just copy them.

(From OE-Core rev: a8fe5fd62fd98e33180f3de700ed01f9efb74a50)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package_manager: Use apt-ftparchive for deb packages

Use apt-ftparchive to create a Release file compatible with SecureApt.

apt-ftparchive is also a more efficient replacement of
dpkg-scanpackages:

root@neopili:~/curro/qtec/qt5022/build-qt5022-cesium/build/tmp/deploy/deb/bobcat
_64# time PSEUDO_UNLOAD=1 apt-ftparchive packages . >/tmp/kkk
real 0m26.873s
user 0m20.968s
sys 0m1.212s

root@neopili:~/curro/qtec/qt5022/build-qt5022-cesium/build/tmp/deploy/deb/bobcat
_64# time PSEUDO_UNLOAD=1 dpkg-scanpackages . >/tmp/kkk
dpkg-scanpackages: info: Wrote 6022 entries to output Packages file.
real 0m59.721s
user 0m16.668s
sys 0m11.164s

apt-ftparchive is not compatible with libpseudo. The calls to ftw()
returns the path in absolute format instead of relative. This
produces
wrong Packages and Release files.

ie:
MD5Sum:
d20227a958f6870137ce0e41b7b84307             1453
/home/ricardo/curro/qtec/qt5022/build-qt5022-cesium/build/tmp/deploy/deb/all/Release

This is why it is called with PSEUDO_UNLOAD.

(From OE-Core rev: c9899a7605f15f7f1ae30c4624d53c7da825b00a)

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

openssh: Update init file to add ED25519 Key generation

[YOCTO #5983]

(From OE-Core rev: c859f9a9f765654a292d57a2bfcea8b4f4c65087)

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

tar: add knob whether acl.h are checked

Previously, it still was checked when there was no sys/acl.h in sysroots directory.
Add knob to decide whether acl.h are checked or not.

Fixed by using PACKAGECONFIG to check acl, with default disabled set.

(From OE-Core rev: ab0bbeeb0b0f6c3c5c7298929cfee757d7bbb111)

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

gettext-0.16.1: kill target m4 macros from sysroot

Ever since the change to how aclocal files are copied (based on dependencies),
target m4 macros seem to more reliably be used in preference to native (which
they should), but in a non-gplv3 build, gettext is 0.16 while gettext-native is
0.18, causing a 0.16 po.m4 to be used with our 0.18 po/Makefile.in.in files,
causing at least some failed builds, including e2fsprogs.

Anyone inheriting gettext will have both gettext-native and gettext available,
and we don't want to use older macros from the target gettext in a non-gplv3
build, so kill them and let dependent recipes rely on gettext-native.

[YOCTO #5964]

(From OE-Core rev: f43139adceaf8039d5347dea0a116dc5923b24c1)

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

pulseaudio: Make it compatible with systemd-209

(From OE-Core rev: 662ff5986370e59cf8638aab66f095f284a7238d)

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

Upgrade to systemd 211+

Fixes [YOCTO #5940]

Fix checks for %ms format to be a link time check
runtime checks wont work in cross compiling

Add a patch to workaround missing _SC_PHYS_PAGES
in uclibc

(From OE-Core rev: 0f5256d4ae5ed88c62e737e3c31587d7635b5dd6)

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

bitbake: runqueue: Revert child signal handler for now

We're running into processes using 100% cpu. It appears theses are locked in
a subprocess.poll() type loop where the process has exited but the code is
looping as its not handling the ECHILD error.

http://bugs.python.org/issue14396
http://bugs.python.org/issue15756

This is likely due to one or both of the above bugs. The question is what actually
grabbed the child exit code as it wasn't this code. Its likely there is therefore
some other code racing and taking that code, it may be some kind of race like:

http://hg.python.org/cpython/rev/767420808a62/

where the fix effectively catches the childs codes in a different part of the system.

We could try and get everyone onto python 2.7.4 where the above bugs are fixed however
for now its safer to admit defeat and go back to polling explictly for our worker exit
codes.

(Bitbake rev: 5b9a099ec2a1dc954b614e12a306595f55b6a99e)

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

bitbake: runqueue: Don't catch all child return codes

Catching all child exit status values is a bad idea. Setting an http sstate mirror
is a great way to view that spectacularly break things. The previous change did
have good code changes so don't revert those parts.

(Bitbake rev: fa7ffb62d510ac1124ae7e08fa4d190a710f5b54)

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

bitbake: runqueue: Really fix sigchld handling

There are several problems. Firstly, a return value of "None" can mean
there is a C signal handler installed so we need to better handle that
case. signal.SIG_DFL is 0 which equates to false so we also need to
handle that by testing explicitly for None.

Finally, the signal handler *must* call waitpid on all child processes
else it will just get called repeatedly, leading to the hanging behaviour
we've been seeing. The solution is to only error for the worker children,
we warn about any other stray children which we'll have to figure out the
sources of in due course.

Hopefully this patch gets things working again properly though.

(Bitbake rev: 973876c706f08735c1b68c791a5a137e5f083dd2)

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

pixbufcache: Fix librsvg-native build

bitbake librsvg-native would fail with an error about missing icu-native.
The reason is that bitbake doesn't directly parse setscene dependencies. This
change ensures bitbake does see the dependencies and avoids the error.

Ideally we'd teach bitbake about those but that is a significant and complex
change so this resolves the problem for now.

[YOCTO #5926]

(From OE-Core rev: 33fa7c8f28d343ecec354a551d45f23643becd59)

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

autotools: Exclude variables from autotools_copy_aclocals

The autotools aclocal copy function should not depend on various variables, these
are accounted for in other parts of the system. Therefore exclude them.

This was causing differences in sstate checksums between different systems and meaning
the sstate cache wasn't being reused as much as it should.

(From OE-Core rev: b7193fadb1a53c86ffe4982a2fa9c1179a74de46)

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

bitbake: runqueue: Ensure handler does not recurse

Failures on the autobuilder look like this handler is recursing. That
shouldn't be possible but it doesn't hurt to code as such.

(Bitbake rev: e39e85803cbe1ef9413a118868c19087c0546d01)

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

bitbake: runqueue: More carefully handle the sigchld handler

We've noticed hanging processes which appear to be looping around
waitpid. Its possible multiple calls to teardown are causing problem
or in theory multiple registrations (although the code should not
allow that). Regardless, put better guards around signal handler
registration.

(Bitbake rev: 79acfb0853aa3215215cee89a945f8e97b0a8fae)

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

bitbake: hob: sync after image deploy

Showing "Deploy image successful" after 'dd' returns
may determine the user to disconnect the usb stick even
though the writing operations are not finished.
This patch makes sure that the entire image is deployed
on the usb stick before the user is informed about any result.

[YOCTO #5892]

(Bitbake rev: cc98b19112ab875ebc7cb604cd96acadac4cbf21)

Signed-off-by: Valentin Popa <valentin.popa@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: runqueue: Don't error if we never setup workers

If we didn't setup any workers (such as bitbake -S), this would error
since we're trying to set a signal handler to None. This patch
avoids that problem.

(Bitbake rev: ce17478c8197abf178c00774f5bbe23fd4375ee2)

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

gettext: upgrade from 0.18.3.1 to 0.18.3.2

Upgrade gettext from 0.18.3.1 to 0.18.3.2.

(From OE-Core rev: b681d264ab2f721c537bfa87cc2ddbbbf3f8c1a2)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

grep: upgrade from 2.16 to 2.18

Upgrade grep from 2.16 to 2.18.

(From OE-Core rev: ac53ef459e1f0e734bdf0a12546675e16631fae8)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sudo: upgrade to 1.8.9p5

Upgrade sudo from 1.8.9p1 to 1.8.9p5.

(From OE-Core rev: afc9e06c4cd9fdb52a751c480d7575a6d611ae34)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

alsa-tools: update to version 1.0.27

Update alsa-tools to latest version 1.0.27:

* remove PR
* update autotools.patch
* add build dependency gkt+3 that hdajackretask requires it.

(From OE-Core rev: 6c699b513c8defe6d4b01ec760185865a29d4582)

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

directfb-examples: update to 1.7.0

Update directfb-examples to latest version 1.7.0.

(From OE-Core rev: ba5e58e2013bfdce7a22e4a56c58fd7c7151cfc6)

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

directfb: update to 1.7.1

Update directfb to 1.7.1:
* remove PR
* remove rename-no-instrument-function-macro.patch which is merged

(From OE-Core rev: 6188b43cdbe4ef0f1a8e9b4345e37058c3f35ead)

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

qmmp: update to 0.7.5

Update qmmp to 0.7.5 version

* update no-host-paths.patch
* remove no-sessionmanager.patch which is merged

(From OE-Core rev: fa10bd5aed84768ded25cdba348ce197a5fdbf2a)

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

docbook-sgml-dtd-native: remove catalog file when do_clean

The docbook_sgml_dtd_sstate_postinst adds catalog file to
/etc/sgml/sgml-docbook.cat and sgml-docbook.bak, but the do_clean would not
remove the file, which would cause unexpected errors since we have multiple
versions of docbook-sgml-dtd:

docbook-sgml-dtd-3.1-native_3.1.bb
docbook-sgml-dtd-4.1-native_4.1.bb
docbook-sgml-dtd-4.5-native.bb

If we run:

$ bitbake docbook-sgml-dtd-4.1-native
$ bitbake docbook-sgml-dtd-4.1-native -ccleansstate && bitbake docbook-sgml-dtd-3.1-native
$ bitbake docbook-utils-native

We would get this error:

jade: cannot open "/path/to/etc/sgml/sgml-docbook-dtd-4.1.cat" (No such file or directory)
make[2]: *** [sgmldiff.html] Error 1

Remove the catalog entry when do_clean will fix the problem.

[YOCTO #5949]

(From OE-Core rev: 87df114982db7a94fce8622b15458bbe2d2ddacc)

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

boost: Pass CFLAGS, CXXFLAGS and LDFLAGS correctly into boost.build

The compileflags and linkflags do_boostconfig puts into user-config.jam are
having no effect. According to
http://www.boost.org/boost-build2/doc/html/bbv2/reference/tools.html the
correct syntax would be <compileflags>"the flags" <linkflags>"the flags".

Since the flags specified were having no effect they can't be necessary and
can be safely removed. We should be passing ${CFLAGS}, ${CXXFLAGS} and
${LDFLAGS} instead so that users of the recipe can pass arbitrary flags as
they choose.

(From OE-Core rev: f09d2583894be0a4069600381046bec5e7f0277b)

Signed-off-by: Mike Crowe <mac@mcrowe.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-core-tools-testapps: added connman tools and tests

Added connman-tools and connman-tests packages.

(From OE-Core rev: 55f596fdfdacd3f5eeabdcabd62545b2d583cc2e)

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

packagegroup-core-tools-testapp: added 3G tools

Added ofono tests package.

(From OE-Core rev: c6180c9dc6466f42b87c2a465c14be3d2e2239cb)

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

apt-native: Install apt-ftparchive

apt-ftparchive is needed to create a Release file compatible with
SecureApt.

It is also a more efficient replacement of dpkg-scanpackages.

(From OE-Core rev: 383e6c7d5fa1f6f02b50155a77d7c82237c11ba9)

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mips64-linux: set ac_cv_sizeof_ssize_t

Set ac_cv_sizeof_ssize_t for mips64;
mips-common will not overwrite it.

"ssize_t is a posix define which is architecture specific whose value
is signed size_t, glibc/uclibc for mips64/n64 linux platform defines
it to be equivalent of 'long' and long here is 8bytes because
mips64/n64 follows LP64 model. In OpenEmbedded our default ABI for
mips64 platforms is N64, having said that autoconf decides to poke at
the platform for finding these kind of sizes which fails when you are
doing cross compiling hence we have to cache it." - Khem Raj

[YOCTO #5935]

(From OE-Core rev: 7a5b6b968c5adf06982ff0f53871f87aaed41580)

Signed-off-by: Valentin Popa <valentin.popa@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

make: add 'inherit pkgconfig'

Bitbaking make-native generates syntax error during
configure: 'PKG_PROG_PKG_CONFIG: command not found'.
Add 'inherit pkgconfig' to solve this issue.

(From OE-Core rev: 841565db49b113da0eec3bfbad934c36bf28ee4b)

Signed-off-by: Valentin Popa <valentin.popa@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto-custom: Use SRCREV_machine

The Yocto kernel tools look for SRCREV_machine in do_validate_branches,
if it's empty, it just returns and silently continues. This likely needs
at least a warning. However, this recipe should be using SRCREV_machine,
and not just SRCREV.

(From OE-Core rev: f36de92551c8c44a15f1997e65cd8ee957143d95)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Reported-by: Saul Wold <sgw@linux.intel.com>
Cc: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe: turn oe into a namespace package

This will let folks extend the oe package with modules from other layers.
Given openembedded consists of more than just oe-core, I think this makes
sense, and adds some useful flexibility.

(From OE-Core rev: 8dfd28925fc47d8a4a1f10ec65df665c8326edc4)

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