uboot-config.bbclass: Skip the package if no valid configuration is found

The ValidError case makes parsing fail even if the U-Boot variant is
not in use for the specific machine and this is not desired. So
instead of raising a parsing error we skip the package.

(From OE-Core rev: d265216dab8146cda17b9ec6167346749896a505)

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

python3: Fix race condition at high parallelism factor

Backport a patch to fix race condition as reported here

https://bugzilla.yoctoproject.org/show_bug.cgi?id=5884

[YOCTO #5884]

(From OE-Core rev: 209a019b12f2941f8aefac9f192e9cdf691196e3)

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

adt_installer: fix issue with x86_64

When the default tune for x86_64 was changed to core2-64, the
environment setup script name did not contain x86 anymore. Hence, the
adt_installer failed for x86_64.

This commit contains a generic fix and is supposed to work with any kind
of machine/tune setting. It's actually extracting the environment script
name using 'opkg files meta-environment-MACHINE'. So, no need to do any
other sort of searches. We know exactly which is the environment setup
script for the specified machine.

[YOCTO #5806]

(From OE-Core rev: a88c386239ddc5816d9045d12cf6db4872fa86da)

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

core-image-minimal-initramfs: Use PACKAGE_INSTALL instead of IMAGE_INSTALL

PACKAGE_INSTALL is set from the IMAGE_INSTALL variable with some additional
items, since PACKAGE_INSTALL is intended to be more an internal variable, use
it instead of the IMAGE_INSTALL which is recommeded for installing additional
packages. This will allow the initramfs recipe to use a fixed set of packages
and not be affected by IMAGE_INSTALL.

[YOCTO #5791]

(From OE-Core rev: abf40223d1412ee8f9d2b5269fad7c6aca6c2570)

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

systemd: sed installed file instead of original

When we change the ROOT_HOME the sed regex does not match if we are running
the do_install() a second time, so copy the units files first to a pre_sed,
so that the next time, we can copy the the original so that the sed regex
is matched correctly in the original

[YOCTO #5765]

(From OE-Core rev: 8a60d490755c2c3010a87f2616008aee2c9cc966)

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

base-passwd: sed installed file instead of original

When we change the ROOT_HOME the sed regex does not match if we are running
the do_install() a second time, so sed the installed file and preserve the original
so that the sed regex is matched correctly in the original

[YOCTO #5765]

(From OE-Core rev: 6007b955ce990e493a9dbf225290a9c7e133feee)

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

base-files: sed installed file instead of original

When we change the ROOT_HOME the sed regex does not match if we are running
the do_install() a second time, so sed the installed file and preserve the original
so that the sed regex is matched correctly in the original

[YOCTO #5765]

(From OE-Core rev: 47d824657acc55e094d5703eed68853f2048c30c)

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

initrdscripts: Add rootimage option

This allows for setting the ROOT_IMAGE name on the kernel command line

[YOCTO #5387]

(From OE-Core rev: 00e3acde7910a5fb1d2e6b71187f2d9283319e71)

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

base: Remove pointless createCopy/update_data calls

I was asked why we had these copy and update_data calls. I can
offer no good explanation other than them perhaps once being needed
for reasons long since forgotten and superceeded.

With modern bitbake these should not be needed and shouldn't have been
for a long time. Lets therefore remove them and simplify the code.

(From OE-Core rev: 33746924bea27e6f4d85898fe37d3e07d4317a3a)

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

security-flags: Avoid lttng-tools issue on arm

(From OE-Core rev: 010d5b437413156c3f4dc90a14698231bb195c2e)

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

security-flags: Deal with powerpc build issues

Building powerpc machines with the standard security flags generated numerous
build failures. Use a reduced set of flags for now to avoid linker issues
and other compile failures.

(From OE-Core rev: 4ef8f658874282ead0c46352474fdb03ad1f1038)

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

security_flags: disable PIE flags for cups builds

(From OE-Core rev: c564bffe7a32470578a22b70e868e7bec2da0a69)

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

runqemu: Use readlink instead of realpath

(From OE-Core rev: 5a4b5c6b8ebd5f8d29888aafcd9608e03717bcd5)

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

dpkg-compiler.m4: remove -Wvla (fix build on CentOS 5.8)

Remove the -Wvla flag from the set of compiler warning flags, since gcc
on old host systems such as CentOS 5.8 doesn't support it, and it
causes a build error for dpkg-native.

(From OE-Core rev: 736ef878570ebe60845da88094907ad28f7b50ff)

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

dpkg: fix a link problem for dpkg-native on CentOS 5.8

[ CQID: WIND00392830 ]

CentOS 5.8 provides the kernel support and headers for the
sync_file_range() syscall, but glibc 2.5 doesn't implement the
sync_file_range() syscall stub, so we can't link dpkg-native. Add a
patch that makes dpkg require a glibc version >= 2.6 in order to use
sync_file_range().

(From OE-Core rev: 197dfda0d971e5e423f1b04a13fbe7ab22d2e874)

Signed-off-by: Donn Seeley <donn.seeley@windriver.com>
Signed-off-by: Lei Liu <lei.liu2@windriver.com>
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

grub: exclude .pc from POTFILES.in (fix build on CentOS 5.8)

Exclude the .pc from po/POTFILES.in since quilt uses "patch --backup",
which will create the backup file under .pc, this may cause unexpected
errors, for example, on CentOS 5.x, if the backup file is null
(newfile), it's mode will be 000, then we will get errors when xgettext
try to read it.

(From OE-Core rev: fe037904ce8a90f428645c19587913c7b90652ef)

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

sqlite: rename to match upstream versioning

Since we are re-doing the version from a big number down to a dot
based number we have to bump the Package Epoch (PE).

(From OE-Core rev: 5ecba3fa5e15e1d329aa1e68349d134db450d383)

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

oeqa/targetcontrol: properly get the host ip

For the SimpleRemote target where we need the host ip and
it wasn't set in conf, we tried to determine it automatically.
However ip route output isn't the same for every network, we
need the last field from the first line.

(From OE-Core rev: 19af3ac53690b1bd2fee54827090f59c525d2236)

Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package_manager.py: create index files for all backends in PACKAGE_CLASSES

The previous implementation was checking IMAGE_PKGTYPE and created the
index files just for the backend used to create the image. Apparently,
'bitbake package-index' should attempt to create the index files for all
backends specified in PACKAGE_CLASSES.

[YOCTO #5827]

(From OE-Core rev: 0521d48a1612bfc735e2c86acc9b685c1dc389ef)

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

bitbake: hob: replace the use of hob-toolchain with populate_sdk task

Now bitbake has an API to run individual task for targets.
Hob can use this to build the sdk, instead of using the hob-
toolchain.

(Bitbake rev: 4a5009036e9cb38f6e0260a88278948931073bc6)

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

meta-hob: remove hob-toolchain since this is not needed anymore

A patch to replace the use of hob-toolchain with bitbake api
was sent to bitbake-devel list and it is needed for this one.

(From OE-Core rev: cb84a4acbd4ad448f140605769b873068e0a0874)

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

sqlite3: Update to 3080301

(From OE-Core rev: 78d9d0ec716fcaaebe5e8361418c3176cf419b95)

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

package_manager.py: correctly handle empty opkg-query-helper.py output

If the output from opkg-query-helper.py is empty, output.split('\n')
would result in a list containing one element which is an empty string
while iterating over each line in the output. An exception is then
thrown by the line:

    pkg, pkg_file, pkg_arch = line.split()

with the message:

    Exception: ValueError: need more than 0 values to unpack

To avoid this, we add a condition to only split the output if it isn't
empty.

(From OE-Core rev: ee7b75c895e77ab20f728423c8efc2ced92265e8)

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

distro/defaultsetup: Add seperatebuildddir.inc

This has been in testing for long enough in various distros and setups,
lets make it the default.

(From OE-Core rev: 875d8d076bf7678321b847425590bbe06765bb84)

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

pango: upgrade to 1.36.2

(From OE-Core rev: 0f16ca1aae1aacd30bde31dd89e9f4fc1d042f36)

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

curl: upgrade to 7.35.0

(From OE-Core rev: 5223646626693a5783919a600fb080a4c6dff06d)

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

sstate: Drop 'SafeDep' code from setscene validation function

I have a feeling this code exists from the time before we had proper
coverage of one sstate task by another task. At that time it was a
"poor" persons version of that idea, we now have much better
code internal to bitbake which handles this.

Worse, this code actually breaks certain rebuild scenarios,
e.g.:

bitbake libtool-cross
bitbake libtool-cross -c cleansstate
rm tmp -rf
bitbake libtool-cross

would fail as binutils-cross wasn't installed from sstate.

The easiest fix is to remove the obsolete/broken code.

[YOCTO #5773]

(From OE-Core rev: ccad07f35fb7f959e24fd50d04c7d10dd5cf20d0)

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

package: Drop do_package_write task

The reasons this task was introduced are lost in the mists of time. It
allowed for the a single "package_write" task instead of spelling out
the explicit package backends, however in all but one case we do that
anyway.

As such as might as well give in and delete the task, converting that
single reference into explicit dependencies.

This gives bitbake a bit less work to to when processing the runqueue
since there are less tasks (but more dependencies in some cases).

(From OE-Core rev: cf70e15f063716f3227d467ab1f4bfc0018286f6)

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

documentation.conf: add a few missing task descriptions

We didn't really need these when the task descriptions were only used
for Toaster, but now we're showing them in -c listtasks they are useful
to have.

(From OE-Core rev: f9d04fcab577d1f48329a4cfe51b1e73fa5d9ba2)

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

classes/utility-tasks: make do_listtasks a little more friendly

* Sort the list so it's at least in some form of logical order. I looked
  at sorting by dependencies, but that's a topological sort, and given
  no such function is shipped as part of the python standard libraries
  it would seem excessive to pull one in just for this. In any case, I'm
  not sure that for the data we have this would lead to any particularly
  pleasing result.
* Show the doc values as defined in documentation.conf (where present)
  as a description

Addresses [YOCTO #4856].

(From OE-Core rev: 36828f8a0db83b5222a8589984e4a02aeb00eada)

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

classes/package_tar: fix conflicts with package_deb / package_ipk

Avoid tar noticing that the directory is changing when
do_package_write_deb or do_package_write_ipk are running at the same
time as do_package_write_tar (because DEBIAN and CONTROL are being added
and removed while tar is running so the directory changes).

Fixes [YOCTO #5652]

(From OE-Core rev: d000761acdb2645ac879d8d9d6b022770545f644)

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

classes/sstate: fix taints being undone on execution of sstate tasks

The code here that deletes stamps was also deleting the taint files; so
forcing an sstate task with -f would force it to execute and then
because the taint file was deleted in the process, the next execution
would simply restore the output from sstate again. We need to exclude
the taint files just like we did in bb.build.make_stamp().

Fixes [YOCTO #5805].

(From OE-Core rev: 4708859e5627488251dc4250d45cb5f4e9736b8a)

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

lib/oe/rootfs: Fix LDCONFIGDEPEND check

We shouldn't be running this if it expands to None or "" or False
so update the code accordingly.

(From OE-Core rev: 255079780729f0dd9df6f35302169e0c8480eced)

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

bitbake: fetch2: Fix mirror repo tarball creation

A typo was meaning that the mirror creation method wasn't being called
when it should have been. Fix the type to fix mirror tarball creation.

[YOCTO #5284]

(Bitbake rev: 348971d410bfd5d8b1757468d73e1d24ae78a594)

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

bitbake: runqueue: Catch ValueError from pickle.loads

* exception like this keeps spinning quite quickly generating GBs of logs
  better to kill it asap and show invalid pickle

(Bitbake rev: a69eb4c12c71bba9d742c4e5578f25c388d9f825)

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

bitbake: ast: Fix support for anonymous methods in wildcard .bbappend files

When using wildcard .bbappend files with anonymous methods in them,
bitbake/python fails to parse the generated code since the '%' is encoded
in the generated method name.

Fix this by including '%' in the convert-to-underscore list during
method name mangling.

While we're at it, move the method name mangling translation table
to a class variable, as suggested by Chris Larson.

[YOCTO #5864]

(Bitbake rev: 537f1f9bbe110acc9848ef95f43468c07d87af79)

Signed-off-by: Jacob Kroon <jacob.kroon@mikrodidakt.se>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lz4: fix CC

It defined the CC to ${TARGET_PREFIX}gcc which lost the --sysroot and was
incorrect, it would cause unexpected errors, we should define it as CC=${CC}.

[YOCTO #5869]

(From OE-Core rev: 5eb0f6e03e5a543f7bad6fcf0cab4173cc8882d8)

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

x246: Disable Altivec asm when SPE is enabled.

asm code is written for Altivec, and should be disabled
for powerpc SPE enabled targets.

(From OE-Core rev: c325b1470c2e009c6b228aa0720bff05452716e4)

Signed-off-by: David Nyström <david.c.nystrom@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image_types.bbclass: fix cpio IMAGE_CMD to preserve working directory

The working directory is changed in a subshell when executing cpio to
preserve the working directory for any subsequent commands. This is to
keep the working directory consistent when generating multiple image
types.

(From OE-Core rev: 5ff6cb920f8be9068a23f7bf0cb1b9a9ff1eda5b)

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

manifest.py: remove redundant import line

The import oe.package_manager line is redundant, let's remove it.

(From OE-Core rev: ade1f3ce34489bba5a7ab23793c1d82559150583)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rootfs.py: tweak _multilib_sanity_test for ipk incremental image generation

The _multilib_sanity_test installs multilib packages in a temporary
root fs, and compare with the current image to figure out duplicated
files that come from different packages.

While incremental image generation enabled and the previous image
was existed, there was an Multilib check error:
...
ERROR: Multilib check error: duplicate files tmp/work/qemux86_64-poky-
linux/core-image-minimal/1.0-r0/multilib/lib32/lib/libc.so.6 tmp/work/
qemux86_64-poky-linux/core-image-minimal/1.0-r0/rootfs/lib/libc.so.6
is not the same
...

The reason is the file in the existing image has been prelinked by
previous image generation and the file in a temporary root fs is not
prelinked, even though both of them came from the same package, the
Multilib check failed.

[YOCTO #1894]
(From OE-Core rev: 8d813f614cdfda31c85bbaf133f2822f90a4a78a)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rootfs.py: support BAD_RECOMMENDATIONS for ipk incremental image generation

While incremental image generation enabled and the previous image is
existed, if BAD_RECOMMENDATIONS is changed, the operation on the
existing image is complicated, so remove the old image in this situation.

The same with PACKAGE_EXCLUDE and NO_RECOMMENDATIONS.

[YOCTO #1894]
(From OE-Core rev: 0566de3fa424af3bdfadcd0a08ce4c214abda083)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rootfs.py: support ipk incremental image generation

The incremental image generation is based on the previous existing
image, adds new packages, upgrades existing packages, and removes unused
packages.

[YOCTO #1894]
(From OE-Core rev: adf587e55c0f9bc74f0bef415273c937401baebb)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package_manager.py: support ipk incremental image generation

Add the following three functions to OpkgPM class:
- The 'dummy_install' is used to dummy install pkgs, and returns the log
  of output;
- The 'backup_packaging_data' is used to back up the current opkg
  database;
- The 'recover_packaging_data' is used to recover the opkg database
  which backed up by the previous image creation;

Tweak 'remove' function in OpkgPM class, which the options for remove
with dependencies was incorrect.

Tweak 'handle_bad_recommendations' function in OpkgPM class:
- Fix none value check;
- Add the existance check of opkg status file;

[YOCTO #1894]
(From OE-Core rev: 5df18065e267a7e55a990ac3728414bb6e28a723)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

manifest.py: add create_full for OpkgManifest class

The function create_full creates the manifest after the package in
initial manifest has been dummy installed. It lists all *to be
installed* packages. There is no real installation, just a test.

[YOCTO #1894]
(From OE-Core rev: 494adecd878496c2edc663ba09a456a9735d8252)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

psmisc: Update to 22.21

(From OE-Core rev: 065862225a5903cd871ba9a7d00bbb307240e9aa)

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

libglade: don't use gnome.bbclass

gnome.bbclass pulls in more functionality than this recipe needs (icon cache,
MIME, GConf), so instead just use gnomebase.bbclass.

[ YOCTO #5842 ]

(From OE-Core rev: f5dff529e5f87b11c5730209df51604181b74321)

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

runqemu: Ensure ROOTFS path is absolute

There is a problem if a relative path is passed to the kernel for NFS usage
that it will not correctly find it, so ensure that the ROOTFS path is absolute.

[YOCTO #2807]

(From OE-Core rev: 5722be0ddda4ec3c96c06b425e5c7e0194326253)

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

insserv: enable ptest support

Install insserv test suite and run it as ptest.

(From OE-Core rev: 5bb786241907bb0304edb25ac1d398b3ae7f3be0)

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>

gtk+3: upgrade to 3.10.7

(*) removed patches that were already
    commited upstream
(*) added new patch to solve the build issue
(*) changed the package URL because
    the old one is unreliable.

(From OE-Core rev: c885365107cc31ba2eff94cfae4f09813ac53a97)

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

git: update to 1.9.0 release

(From OE-Core rev: 2d51215dbfcde7a9ead430ca8c127944c77ceb40)

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

openssh: enable PAM at runtime based on DISTRO_FEATURES

Everything is setup to use PAM except for the server config. If 'pam' is
in DISTRO_FEATURES the configs will be changed to enable it.

Syslog will now show:

Feb 19 09:28:36 beast sshd[2980]: pam_unix(sshd:session): session opened for user koen by (uid=0)

And more importantly:

koen@beast:~$ loginctl
   SESSION        UID USER             SEAT
        c1       1000 koen             seat0
        c3       1000 koen             seat0
       c13       1000 koen

3 sessions listed.

Systemd now registers the session properly so it won't kill things like 'screen'
and 'tmux' when disconnecting the ssh session.

(From OE-Core rev: 08c523daec0f5161803d2a6704402490c18ff6ca)

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

update-rc.d.bbclass: fix inhibit check

In update-rc.d.bbclass it checks variable INHIBIT_UPDATERCD_BBCLASS to
inhibit from inheriting this class. But it is wrong logic that when
'sysvinit' is in 'DISTRO_FEATURES', INHIBIT_UPDATERCD_BBCLASS will not
be checked.

Replace 'or' with 'and' to fix it.

(From OE-Core rev: a0353102184892f09d3f97f018e457a4684550ea)

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

uclibc: Add new functionality needed for systemd 209

implement eventfd_read and eventfd_write and setns

Define F_SETPIPE_SZ and F_GETPIPE_SZ

(From OE-Core rev: ef0adae63b4b9689c25ed7d84f2b09575c81ce83)

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

valgrind: integration of regression tests to ptest

Modifies valgrind's regression test framework to be compatible
with the yocto PTEST framework as follows:
 * existing recipe valgrind*bb adds new methods: do_compile_ptest and
   do_install_ptest.
 * new file run-ptest adds the wrapper interface to the valgrind
   regression test script vg_regtest.
 * existing valgrind regression test script 'vg_regtest' changes
   to report the status of the valgrind component tests in the
   format that PTEST expects, instead of the valgrind formats, but only
   when vg_regtest is invoked with an optional --yocto-ptest argument
 * four new patches disable building tests that don't compile with
   the yocto compiler and default options. See the patches for details.

(From OE-Core rev: d4438e421f448cdb7e25c038d657bbebc1b6486e)

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

wic: Make exec_native_command() fail if a command isn't found

Because exec_cmd() return values can in certain cases be non-zero yet
non-fatal, we don't want to automatically make them fatal (though
there should at least be a warning in such cases, which this patch
also does); non-zero return values are definitely fatal however if
they mean that a native command wasn't found, so have
exec_native_cmd() check the return value of exec_cmd() for that case,
and bail out if so.

[YOCTO #5835]

(From OE-Core rev: 43ac6e3216c5d985d6f90a28e727e397df616267)

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

wic: Fix exec_native_cmd() path

The path exported in exec_native_cmd() includes bogus 'PATH=' which
means the native paths for all but the first will be ignored.

(From OE-Core rev: e10c1102d958ffc6b521fb36c6bc51d896503218)

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

image-vmdk.bbclass: Don't force syslinux timeout

Don't force SYSLINUX_TIMEOUT and SYSLINUX_PROMPT if overrides have
been defined elsewhere for an image configuration.

(From OE-Core rev: a6522c19af9525a6e0bed241f2d926faaef76827)

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

opkg: Upgrade to v0.2.1

v0.2.1 of opkg is a minor bugfix release.

Tested on qemux86 to ensure opkg-native can build an image and opkg can update,
remove and install packages.

(From OE-Core rev: 935abb7f867944dd4f169469045771d0600bbbd2)

Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

recipes: convert remaining SUMMARY/DESCRIPTION cosmetic issues (part 2)

Completes previous commit b5292d4115a4555a66b5e35acdc67dd71fb8577f.
Updates SUMMARY[doc] (meta/conf/documentation.conf).

Changes:
- rename DESCRIPTION with length < 80 to (non present tag) SUMMARY
- drop final point character at the end of SUMMARY string
- remove trailing whitespace of SUMMARY line

Note: don't bump PR

(From OE-Core rev: ad17dfd31a2b97b3e610a0ea0889f5ecb2a63b97)

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

package_manager.py: Fix $i -> %s in bb.note

* change the text a bit, because I have trouble parsing two
  verbs at the beginning of sentence.

(From OE-Core rev: 9837d194682f46e3a0363765bd6b317180a4cda0)

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

libpcre: enable ptest support

Install libpcre test suite and run it as ptest.

(From OE-Core rev: fadaa1c7f4d927e3003bc4750f3b677ed627febb)

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>

opkg: Fix add-exclude.patch

The case statement for ARGS_OPT_ADD_EXCLUDE added to the argument handling
switch statement in opkg was missing a "break;" at the end, so it was falling
through into the handler for ARGS_OPT_NOACTION. Thus when "--add-exclude" was
specified on the command line it was as if "--noaction" was also being
specified. This appears to be the root cause of YP bug 5311.

Tested using the case described by Alexandru Georgescu in YP bug 5311:

    MACHINE ??= "qemux86"
    IMAGE_INSTALL_append = " man"
    PACKAGE_EXCLUDE = "man"
    PACKAGE_CLASSES ?= "package_ipk"

Built image and boot tested on qemu, ensured that man was not installed but the
rest of the system was installed correctly.

(From OE-Core rev: c269cea3117d7fa120c02bdb47d1adf18a1bcba1)

Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Cc: Mark Hatle <mark.hatle@windriver.com>
Cc: Saul Wold <saul.wold@intel.com>
Cc: Alexandru Georgescu <alexandru.c.georgescu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gstreamer1.0-plugins-good: removed dependency on gconf

gconf is no longer used as a configuration system.

(From OE-Core rev: 1f8d944860cf1b1c3f4a6c48e4e1bdc9b09ede12)

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

gstreamer1.0-libav: upgrade to 1.2.3

- bugfix release.

(From OE-Core rev: d4c553d7b883e6b1ba7fb931b9df653e7842da05)

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

gstreamer1.0-plugins-bad: upgrade to 1.2.3

- bugfix release.

(From OE-Core rev: 3bfc8c2ee4334d0e8217dc0a6a0a0463051c05f1)

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

gstreamer1.0-plugins-ugly: upgrade to 1.2.3

- bugfix release.

(From OE-Core rev: 717eef1932098db4ac74966d3036f002f0e71210)

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

gstreamer1.0-plugins-good: upgrade to 1.2.3

- bugfix release.

(From OE-Core rev: d1553b6a5ffcc8964098e33130d02ff6a287891c)

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

gstreamer1.0-plugins-base: upgrade to 1.2.3

- bugfix release.

(From OE-Core rev: f841837926bfbfc1a5d61465f9201621b213d2be)

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

gstreamer1.0: upgrade to 1.2.3

- bugfix release.

(From OE-Core rev: 2576324d961334f63ba3fb975bab3088205ea21b)

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

libvorbis: upgrade to 1.3.4

- Updated website;
- Cleaned up bugtracker web address;
- Removed PR;
- Switched to ${BP} variable;
- removed obsolete_automake_macros.patch,
included in upstream;

(From OE-Core rev: 8d0f6b39839d14dea5ed5cb8888466ed76afee00)

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

populate_sdk_*.bbclass: remove left over bash routines

The list_installed_packages bash routine is no longer needed here. We've
got a python replacement.

(From OE-Core rev: 6c1e379ce81a6f397bf8e7e7c9dd3cea91274b46)

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

buildhistory.bbclass: fix creation of installed packages list

Call the new python routines to create the packages list.

[YOCTO #5831]

(From OE-Core rev: e8bd5b85f144a3260e30f67c3d231b324967be3f)

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

package_manager.py: fix installed package list creation for Opkg/Dpkg

Small error in the package list creation routine. Buildhistory was
supposed to use this but was never called. Hence, it escaped tests...

(From OE-Core rev: eaa1994ad22730cec8e8c57736915da6b45a416e)

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

lib/oe/image.py: add image dependency mechanism

This commit adds a dependency mechanism to image creation, so that we can
split the images creation execution in groups, that can be executed in
parallel, having the dependencies satisfied in the same time. The old
code didn't need this since everything was serialized.

Technically, it adds a dependency graph topological sort class that the
main Image class can use to sort out the dependencies.

Images that have dependencies have to declare them using the NEW
IMAGE_TYPEDEP variable, like in the example below:

For:
IMAGE_FSTYPES = "i1 i2 i3 i4 i5"
IMAGE_TYPEDEP_i4 = "i2"
IMAGE_TYPEDEP_i5 = "i6 i4"
IMAGE_TYPEDEP_i6 = "i7"
IMAGE_TYPEDEP_i7 = "i2"

We'll get the following image groups, sorted out by their dependencies:

[['i1', 'i3', 'i2'], ['i4', 'i7'], ['i6'], ['i5']]

The algorithm can probably be optimized but, given the small size of the
graphs, it'll do.

[YOCTO #5830]

(From OE-Core rev: db9dd4b4ef9120baccbccae77d9c31f54a6eb9a1)

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

image*.bbclass, bootimg.bbclass: add image type dependencies

The following dependencies were manually added in the image creation
code. However, in order to have an image dependency mechanism in place,
use a new variable, IMAGE_TYPEDEP, to declare that an image type depends
on another being already created.

The following dependencies are added by this commit:

elf -> cpio.gz
live -> ext3
vmdk -> ext3
iso -> ext3
hddimg -> ext3

This commit adds also another new variable: IMAGE_TYPES_MASKED. Currently,
masking out certain types from IMAGE_FSTYPES was hardcoded in the image
creation code.

[YOCTO #5830]

(From OE-Core rev: 5e2796aa28e02ae3a076c6593c6533753720b13d)

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

avahi: handle SO_REUSEPORT not being available

Linux < 3.9 doesn't have the SO_REUSEPORT option so instead of failing to start
when built with >=3.9 kernel headers but booted on <3.9 kernels, continue as if
SO_REUSEPORT wasn't available.

(From OE-Core rev: 85e89da55f778ad3713460cb0df1435d82e94510)

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

mesa: upgrade to 9.2.5

tested on qemux86, genericx86 for core-image-sato
and core-image-weston.

(From OE-Core rev: a5ebdb6ad8e4f94ac819275d55575230e057e4ae)

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

libxkbcommon: Add PACKAGECONFIG for x11

* allow to explicitly disable x11 with --disable-x11, otherwise
  do_configure fails for DISTROs without x11 in DISTRO_FEATURES:
  | No package 'xcb-xkb' found
  | configure:18763: $? = 1
  | configure:18777: result: no
  | No package 'xcb' found
  | No package 'xcb-xkb' found
  | configure:18793: error: xkbcommon-x11 requires xcb-xkb >= 1.10 which
    was not found. You can disable X11 support with --disable-x11.

(From OE-Core rev: e13741e52cc6b59757942a05c7847fe7fc090f63)

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

image_types: minor, inline CPIO_TOUCH_INIT()

Since generic compression solution has been introduced
(see revision b7e4ed41ee480f00b7265341e9e2d2c2b9135143),
CPIO_TOUCH_INIT() is only called by IMAGE_CMD_cpio.

(From OE-Core rev: e856e04e6134ce98e89027864f69b2ab48d40dcc)

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

cml1.bbclass: Add diffconfig task to cml1

diffconfig() is a new task that makes a diff between the
old and new config files and writes to the fragment.cfg result file.
menuconfig() always copy the original config file, so the user
doesn't need to copy it.

(From OE-Core rev: f0dcbbdcc37da12f84f730b8235f50be8c150c50)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel-yocto: Add do_diffconfig task

Instead of using 'diff' command between two kernel config files,
the task diffconfig does the job creating the file
$WORKDIR/fragment.cfg that user should review and use.

[YOCTO #3862]

(From OE-Core rev: db2a44c8f08f2371a52ff6662d6bc64bc42ad551)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lighttpd: introduce /etc/lighttpd.d subdir for config file includes

Extend default config file by a directive to include config file
fragments from /etc/lighttpd.d. This allows other web application
packages to put their configuration there.

(From OE-Core rev: 949ef58cf0684147b07745bd1199014ac57b437c)

Signed-off-by: Steffen Sledz <sledz@dresearch-fe.de>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

runqemu-export-rootfs: Fix missing # in comment

(From OE-Core rev: 38785af0ca0b020574b32b9e3b21129469a35c94)

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

image_types.bbclass: Fix tar IMAGE_CMD to not change directories

below commit

commit 9d9bca8785911e8ae06d507bbfb99d6a811f072e
Author: Jonathan Liu <net147@gmail.com>
Date:   Mon Feb 17 21:46:28 2014 +1100

    lib/oe/image.py: fix working directory

    The working directory needs to be changed before the image creation
    commands instead of afterwards.

exposed an issue when generating tar images, where it would cd into
rootfs and then create the rootfs tarball and remain there since the
cmd to cd into deploy dir moved up the immediate following cmd like
tar.bz2 or tar.gz were still looking for tarball in current directory
which esentially was not deploy dir but IMAGE_ROOTFS instead

(From OE-Core rev: 9a98b8bfe13b94d89a9e39cca821f26fdeaeffc3)

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

uclibc: Update to git tip

(From OE-Core rev: 4c280911373abb8021b1fea6e136df8b4cabdea3)

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

binutils: ld-is-gold should not affect native and crosssdk recipes

gold linker does not support all architectures, currently arm and x86
are best supported, therefore we can not enable this as a full distro
option where we need to support other architectures e.g. ppc, mips
currently, if we enable ld-is-gold distro feature conditionally then it
invalidates native and sdk version of native binutils because configure
option would change. With this patch we limit ld-is-gold feature to
imapact cross binutils and target binutils only. This means that we
can reuse the sdk and native versions across architectures.

(From OE-Core rev: 542609b968f0dc6f56cc9d6aabe4ff0144041261)

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

icu: Disable the default LDFLAGSICUDT for target compile

By default LDFLAGSICUDT=-nodefaultlibs -nostdlib for Linux
which means DT_NEEDED section for libicu will not be populated
with dependencies when we reset it to be empty then the default
libraries it needs are added to DT_NEEDED section in ELF header

This patch is a workaround as I consider it for now, since it could
be a problem in glibc dynamic loader for arm (especially for hf case)
where its unable to load shared objects which dont have any dependencies
expressed in DT_NEEDED segment.

here is when LDFLAGSICUDT=-nodefaultlibs -nostdlib

Dynamic section at offset 0x1549c10 contains 8 entries:
  Tag        Type                         Name/Value
 0x0000000e (SONAME)                     Library soname:
[libicudata.so.51]
 0x00000010 (SYMBOLIC)                   0x0
 0x6ffffef5 (GNU_HASH)                   0xf8
 0x00000005 (STRTAB)                     0x188
 0x00000006 (SYMTAB)                     0x138
 0x0000000a (STRSZ)                      54 (bytes)
 0x0000000b (SYMENT)                     16 (bytes)
 0x00000000 (NULL)                       0x0

here is one with LDFLAGSICUDT empty

Dynamic section at offset 0x154a014 contains 22 entries:
  Tag        Type                         Name/Value
 0x00000001 (NEEDED)                     Shared library: [libc.so.6]
 0x0000000e (SONAME)                     Library soname:
[libicudata.so.51]
 0x00000010 (SYMBOLIC)                   0x0
 0x0000000c (INIT)                       0x33c
 0x0000000d (FINI)                       0x500
 0x6ffffef5 (GNU_HASH)                   0xf8
 0x00000005 (STRTAB)                     0x204
 0x00000006 (SYMTAB)                     0x144
 0x0000000a (STRSZ)                      192 (bytes)
 0x0000000b (SYMENT)                     16 (bytes)
 0x00000003 (PLTGOT)                     0x154a0f8
 0x00000002 (PLTRELSZ)                   16 (bytes)
 0x00000014 (PLTREL)                     REL
 0x00000017 (JMPREL)                     0x32c
 0x00000011 (REL)                        0x2fc
 0x00000012 (RELSZ)                      48 (bytes)
 0x00000013 (RELENT)                     8 (bytes)
 0x6ffffffe (VERNEED)                    0x2dc
 0x6fffffff (VERNEEDNUM)                 1
 0x6ffffff0 (VERSYM)                     0x2c4
 0x6ffffffa (RELCOUNT)                   1
 0x00000000 (NULL)                       0x0

btw. ldd reveals the fist one to be static library while the second
one is detected as shared library. This could be a clue into how
elf headers are being interpreted by dynamic loader.

The data seems to be all static in libicudata which could
load it quicker and thats what could be confusing dynamic linker ..may
be

(From OE-Core rev: 644c307f63f0f0b7e97140850d3d1d2124b11b1b)

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

rootfs.py: Check for LDCONFIGDEPEND being empty string

We override LDCONFIGDEPEND to be empty string for uclibc
however the current check is for it being None as a result
the function is still executed but ldconfig-native is not
built as dependency for rootfs when building with uclibc

Fixes errors like below

File:
'/home/kraj/work/angstrom-repo/sources/openembedded-core/meta/lib/oe/rootfs.py',
lineno: 191, function: _run_ldconfig
     0187:    def _run_ldconfig(self):
     0188:        if self.d.getVar('LDCONFIGDEPEND', True) is not None:
     0189:            bb.note("Executing: ldconfig -r" +
self.image_rootfs + "-c new -v")
     0190:            self._exec_shell_cmd(['ldconfig', '-r',
self.image_rootfs, '-c',
 *** 0191:                                  'new', '-v'])

(From OE-Core rev: 315a71e00928fea6715087b1ce66f8a8a91d5fe0)

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

eglibc: Upgrade from 2.18 -> 2.19

License formatting and address for FSF in the COPYING and COPYING.LIB
has changed.

Dropped patched already upstream and patches that were workarounds for
older glibc and busybox

for e500 we have should pass --without-fp to eglibc/glibc 2.19 onwards
the code is merged from eglibc into glibc upstream under nofpu/ pretext

(From OE-Core rev: 875df27e56b82fcf970410b6d78e3672471c336a)

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

valgrind: Add glibc 2.19 awareness

It wont compile otherwise and fail with errors e.g.

Valgrind requires glibc version 2.2 - 2.17 ...

(From OE-Core rev: cc90cbc6d36b9f15a461b877b44c7498c2061744)

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

Revert "lib/oe/image.py: fix working directory"

This reverts commit 3f49597225a58965124503ca5f3cc4011b04b3c0.

This change appears to cause more problems than it fixes since the
compression commands usually work in the deploy dir but the archive ones
have always worked in the rootfs dir (which is clear from the tar command
we use).

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

bitbake: toaster: implement package summary page

Implement the updated design for the package summay page, with pop-up
boxes for the layer commit ids, column filtering, and column sorting
support.

[YOCTO #4318]

(Bitbake rev: c39b99792547b642570ea5152070e7396e812390)

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

bitbake: toaster: implement recipe summary page

Implement the updated design for the recipe summay page, with pop-up
boxes for the dependecies and layer commit ids, column filtering, and
column sorting support.

[YOCTO #4294]

(Bitbake rev: 92b71c8e6cfcd656866fbf9bd00bf184b223c5fa)

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

bitbake: toaster: Make popovers mutually exclusive

Table popovers should be mutually exclusive (only one
of them should be open at any given time), and should
close when you click outside their area. But this is not
the default popover behaviour in Bootstrap, so some
additional javascript is needed.

The code in main.js taking care of this in the design
prototype was quite ugly and didn't get on well with
certain browsers. I have replaced it with a better
solution (although still not ideal).

(Bitbake rev: d56633c00f6730c053f355570211eba1bdc41b62)

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

bitbake: toaster: Make popovers scrollable

Some recipes and packages have a lot of dependencies.
To stop their popovers from taking over the full
height of the screen, I have given them a maximum
height and made their content scrollable in
default.css

(Bitbake rev: b6416a98f441516100d2ce7baca30f09714a1d1d)

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

bitbake: toaster: Make all targets links to the build dashboard

Change the build.html template to remove the conditional check
that added links to the build dashboard only to those targets
that are image recipes.

All targets should be links to their corresponding build
dashboard.

(Bitbake rev: 05f52b9fad597173d813fa4305af65f1c9bb88a1)

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

bitbake: toaster: Select a radio button by default in Filter Dialog

If none of the filter options in the filter dialog are active,
then select the first radio button.

(Bitbake rev: 8aa63143cc446227c69f64688b314c65b74604d8)

Signed-off-by: Ravi Chintakunta <ravi.chintakunta@timesys.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: Fix for task_color tag

This patch changes the task_color tag in
projecttags.py to make sure it adds the error
class to the rows corresponding to failed tasks
in the tasks table.

(Bitbake rev: 44c5b3a93a93a464429e8d998aace25f840ae724)

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

bitbake: toaster: Apply error style to cell links

This patch changes default.css so that any anchor
tag inside a table cell with the error class applied
inherits the class styles. This ensures that the
failed tasks listed in the builds table look red
like the failed tasks in the tasks table.

(Bitbake rev: e5cbb190317dba2eddd933678474569563ec840b)

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