bitbake/utils.py: Ensure utils.which() returns full paths

If the path passed to which contains empty elements, it will search
the current working directory for the file which is correct baheviour.

Various pieces of code assume the path returned is a full path though.
This commit ensures we don't return relative paths.

(Bitbake rev: 4de24ccc10e40cc088b8515095df59f69b12715d)

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

bitbake/cooker: Ensure matchFile returns full pathnames

We should always be passing full pathnames around within bitbake. If
a file was referenced as a relative path to the current working
directory, it might not get passed through the abspath call and
hence the cwd would not get added as a prefix.

This change adds a second pass at ensuring we only return absolute paths.

(Bitbake rev: 72fc62ca124a24e2dbe404a3c83a49608a7c7931)

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

Hob:fixed the issue of primary button is not an 'orange hob button'

In image detail page, the primary button should be an 'orange hob
button', there has a primary button is not, so change it.

[YOCTO #2326]

(Bitbake rev: 884a0b2d927404991b7e23d5bbfab096a5b7e849)

Signed-off-by: Liming An <limingx.l.an@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

BBHandler: Tweak some data store access methods to the preferred syntax

(Bitbake rev: d7ec0f90354e60c415a01429bf26746c1d63b7ad)

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

qemu: disable vnc-jpeg compression

This is disabled by default to ensure a deterministic build, qemu
configure would sometimes detect the libjpeg and then enable this.
Since mostly qemu is used locally disable by default.

(From OE-Core rev: 0f795f81fe5ad3ef78c21a177eca90fbc8810f81)

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

qemu.inc: Remove mesa-xlib as PREFERRED_PROVIDER

The xserver-xorg uses and depends on mesa-dri, so we should
use the default PREFERRED_PROVIDER of libgl as mesa-dri.

This resolves the following:
ERROR: Multiple .bb files are due to be built which each provide virtual/libgl (/intel/poky/distro/meta/recipes-graphics/mesa/mesa-dri_7.11.bb /intel/poky/distro/meta/recipes-graphics/mesa/mesa-xlib_7.11.bb).
 This usually means one provides something the other doesn't and should.

(From OE-Core rev: cf8b4b95c6d84c097f4fc63662f181a59e0f9cb9)

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

xserver-xf86-config: Add xorg.conf for arm,mips,ppc,sh qemu machines

(From OE-Core rev: db0defb1d019a0e7c06d37dacc09842174e2f548)

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

qemumachines: Enable xserver-xorg as default xserver

For qemux86 and qemux86-64 include qemu.inc after defining XSERVER

XSERVER variable is also weakly defined in task-core-x11.bb
which means we can not use ??= otherwise when building any qemu image
that uses task-core-x11.bb will get the wrong definition

So we define the XSERVER common set for qemu in qemu.inc
and as we know x86 and x86-64 qemu overrides the default
we include qemu.inc after that definition which means that
qemux86 and qemux86-64 get their own definitions and other
qemus get the definitions from qemu.inc. other non-qemu machine
will get their defintion from task which points to kdrive
as of now.

(From OE-Core rev: 62dba36166bc5faa32ba3e0664ae98b168cde6b1)

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

Update to distro_tracking_fields.inc

The fields for the following packages were updated:

libdrm
xf86-video-vmware
xf86-input-mouse
xf86-video-vesa
xf86-video-intel
xf86-input-synaptics
mdadm
xextproto
xcmiscproto
scrnsaverproto
recordproto
xproto
kbproto
fontsproto
libxt
xtrans
libpciaccess
fontconfig
xcb-proto

(From OE-Core rev: 801bfa7edfbff1522964efab0c0ba40bbc52c601)

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

xf86-input-synaptics: upgrade to 1.6.1

(From OE-Core rev: 0a1e60410234f0315b617593859b9deab89c2cd4)

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

xf86-video-intel: upgrade to 2.19.0

(From OE-Core rev: e4db6a5f86b0e725c0ddbd114316d488a024b951)

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

xf86-video-vesa: upgrade to 2.3.1

Also, licence checksum has changed because a new copyright line from
Red Hat has been added. The rest of licence file, however, is unchanged.

(From OE-Core rev: 21dd5de819251a611a8fe206b2d4bc390008c4f5)

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

xf86-input-mouse: upgrade to 1.7.2

(From OE-Core rev: 44cb88f9acde3ef264a04311ff6c8c195ac9cd57)

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

xtrans: upgrade to 1.2.7

(From OE-Core rev: c2943924d94c2418736b9e935b4b058e882b4752)

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

libxt: upgrade to 1.1.3

(From OE-Core rev: 438705407d9af2b54c03648b23bb16411424d870)

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

xcb-proto: upgrade to 1.7.1

(From OE-Core rev: 03b42a5981c98cc7c47b376ce91e46098c4b8eed)

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

recordproto: upgrade to 1.14.2

(From OE-Core rev: b3d94113bb277b58dad9de021313bafa57c9780b)

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

scrnsaverproto: upgrade to 1.2.2

(From OE-Core rev: 1b134f5240588c53e82effb1cc77ba91edabbcba)

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

xcmiscproto: upgrade to 1.2.2

(From OE-Core rev: 481bc016c4b029200c75feb5a91bd82a425810e3)

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

xextproto: upgrade to 7.2.1

(From OE-Core rev: 02f6d0eede83208b48759319be2b5af99e15af59)

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

kbproto: upgrade to 1.0.6

Also, shifted up one line the start and end lines for licence checksum
computation.

(From OE-Core rev: b34855ef05fc5373f55ea75efa617316016f9d83)

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

xproto: upgrade to 7.0.23

(From OE-Core rev: b7cd93644aaa50aac39ab23c0a2ca19415ba3989)

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

libdrm: upgrade to 2.4.34

(From OE-Core rev: af3c8de7b6cbedddaf6bf89108879c89a52ff02e)

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

mdadm: upgrade to 3.2.5

(From OE-Core rev: 511b25110a5c0948f2be289daf0bf25f32466487)

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

valgrind: Enable valgrind for armv7

Valgrind supports the armv7 architecture, this patch allows armv7 users to build and use Valgrind

This patch was run-tested on a Gumstix Overo (armv7a cortex-a8)
* The test consisted of running valgrinds memcheck (memory leakage detection),
* and callgrind (profiling) on a Qt 4 application

(From OE-Core rev: 211ac8b39023d4998fb7d5a577b63a4960b08172)

Signed-off-by: Samuel Stirtzel <s.stirtzel@googlemail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gawk: update to 4.0.1

(From OE-Core rev: 345566c448f5cd628866cb0736d567f5a610b3db)

Signed-off-by: Radu Moisan <radu.moisan@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

jpeg: update to upstream version 8d

* fix_for_automake_1.11.2.patch removed as issue resolved upstream
* Have README as license file instead of several source files that just
  refer to it

(From OE-Core rev: 330173c75426aa38919091da9ae8a9e90de9e0e2)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glib-2.0: upgrade to 2.32.3

(From OE-Core rev: 1f73485a143f0aa6cd87636f3d36a7d1dfc40a3b)

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

sanity.bbclass: check sanity at BuildStarted rather than ConfigParsed

This enables a user to use bitbake -e even when the sanity checks are
failing.

(From OE-Core rev: be317df78535120137ecaadf797e3f4dfe04119e)

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

sanity.bbclass: catch an extra exception in check_create_long_filename

The call to bb.mkdirhier() in check_create_long_filename() can fail with an
OSError, explicitly catch this and report something useful to the user.

(From OE-Core rev: b066906477eb0496a2babb3d8e87682a1b7df0de)

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

sanity.bbclass: add extra information when SSTATE_CACHE unusable

If the user does not have write permissions to SSTATE_CACHE, detected by
the check_create_long_filename() test failing with a "Permission denied"
value in strerror, then suggest they might want to use the location as
an entry in SSTATE_MIRRORS.

(From OE-Core rev: 719d44305508c75a9f2decacdff6558ca14277bb)

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

sanity.bbclass: add newline to check_create_long_filename failure message

Each failure in the sanity message should be reported on a new line.

(From OE-Core rev: 9362702aa9dd82fed7e6d6e3cb1289521b802075)

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

sanity.bbclass: data.getVar(VAR, obj, exp) -> obj.getVar(VAR, exp)

Replace calls to data.getVar(VARIABLE, data_object, expand) to
direct calls to the getVar method the the data_object.

(From OE-Core rev: 41ee978b62a9c40f36f8ad0acef147e36edffa17)

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

sanity.bbclass: copy the data store and finalise before running checks

At the ConfigParsed event the datastore has yet to be finalised and thus
appends and overrides have not been set.
To ensure the sanity check is being run against the configuration values
the user has set call finalize() on a copy of the datastore and pass that
for all sanity checks.

(From OE-Core rev: 71142172efc0f44a50216550c2b6cc3094fdc21d)

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

Revert "sanity.bbclass: check user can read and write to SSTATE_DIR"

This has caused problems for several users, including the Yocto Project
autobuilder. Since the message was added in order to be more user friendly
revert the change.

This reverts commit 0c0c4efbf92bcf0f8942f17c18525a4b4ed1798c.

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

python: bump PR of packages after update of distutils.bbclass

* Bump every recipe inheriting distutils or setuptools and
  not overriding do_install without calling distutils_do_install.

(From OE-Core rev: 61487342764684cacef765e221609fe16ffc2740)

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

distutils.bbclass: don't delete .pyo files

* Deleting .pyo files causes them to get compiled on the target.
* First boot gets *really* slow for python based projects.
* No space gets saved on the target.
* The package manager doesn't know about the files and
  therefore fails to uninstall them, occupying space and causing
  uninstalled python scripts to remain executable.
* It's inconsistent, because python itself and autotools based
  projects already ship .pyo files.
* Probably .pyo files were deleted because .pyc files were
  available earlier, but this has changed and OE-Core's python
  now only generates optimized .pyo files. Deletion of .pyo was
  introduced in 2008, python/04-default-is-optimized.patch
  was introduced in 2009.

(From OE-Core rev: 25e186ad5b75bd2f93435857580bd16698e18e21)

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

netbase: merge init script updates from upstream

* Read /proc/mounts only once.
* Support many more network filesystem types.
* Remaining differences to netbase 4.47:
  - Uses the mountvirtfs keyword instead of mountkernfs
  - Doesn't use lsb functions
  - Doesn't print a warning if /etc/network/options exists
  - Doesn't use --exclude parameter for ifup, because
    busybox doesn't support it.

(From OE-Core rev: faca42ee5249cf6aae9e53e44bb404de21dd4471)

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wpa-supplicant: fix packaging error since last commit

* wpa-passphrase has its own package, but commit
  4a4c568e25a08e9f222d723f9819582c9f895c58 broke it.

(From OE-Core rev: 7e4fcbb1ceac736c44bffc834f7e8f2b34ac4402)

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libnss-mdns: fix postinst scripts

* On upgrade, postinst ocassionally returned 1, so use a
  conditional instead of &&.
* Use sed patterns in order to make it work more generally.

(From OE-Core rev: c678d4c42ed0075572c5fad17ca33773ad3f6447)

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

ppp: avoid host contamination

* Import patch from OE-Classic commit f7c547ede1bfc94c24f6315a23ff48a4c5b9ffaf.
* Fixes linking with libpcap (ppp filter).

(From OE-Core rev: ba1539f165470d6ee9b7b95b5067792e384e6aea)

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libtool: fixed parallel build related race

While building libtool, the libtool script itself will be regenerated
because OE modifies a dependency[1]. With -jX, this operation (-->
removal, creation of non-x file, 'chmod a+x') can happen at a time when
the script is going to be executed.  This can cause errors like:

| arm-linux-gnueabi-libtool: compile:  ccache arm-linux-gnueabi-gcc ...
| ...
| /bin/sh ./config.status libtool
| ...
| arm-linux-gnueabi-libtool: compile:  ccache arm-linux-gnueabi-gcc ...
| /bin/sh: ./arm-linux-gnueabi-libtool: Permission denied
| make[2]: *** [libltdl/libltdl_libltdl_la-lt__alloc.lo] Error 126

I am not sure whether the custom do_compile_prepend() is still needed.
For now only the issue above will be fixed by executing ./config.status
yet again.

[1] see 648290d5bf4d6ff50d3643bb7ad902dfc23aa702

(From OE-Core rev: e065c11b6066f8c1d37496b58dd853a684cd9136)

Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

ncurses: Avoid occasional builling failure when having parallel processable task

ncurses failure non-gplv3 build (race issue) like the following \
error information:

| tic: error while loading shared libraries: /srv/home/pokybuild \
/yocto-autobuilder/yocto-slave/nightly-non-gpl3/build/build/tmp/\
work/x86_64-linux/ncurses-native-5.9-r8.1/ncurses-5.9/narrowc/lib\
/libtinfo.so.5: file too short
| ? tic could not build /srv/home/pokybuild/yocto-autobuilder/\
yocto-slave/nightly-non-gpl3/build/build/tmp/work/x86_64-linux/\
ncurses-native-5.9-r8.1/image/srv/home/pokybuild/yocto-autobuilder\
/yocto-slave/nightly-non-gpl3/build/build/tmp/sysroots/x86_64-linux\
/usr/share/terminfo
| make[1]: *** [install.data] Error 1

This is a race issue which is caused by
install.libs and install.data:

1) install.data needs run tic
2) tic needs libtinfo.so
3) install.libs would regenerate libtinfo.so
4) but install.data doesn't depend on install.libs, and they can run
   parallelly

So there would be errors in a very critical condition: tic is begining
to run at the same time when install.libs is generating libtinfo.so, and
this libtinfo.so is not integrity, then there would be the  above error.

Let task install.libs run before install.data for fixing this bug.

[YOCTO #2298]

(From OE-Core rev: 80e4833782edc5fbda2a7f5d003a854f127137ec)

Signed-off-by: Xiaofeng Yan <xiaofeng.yan@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

eglibc: package mtrace separately

add libc-mtrace as dependency for task-core-tools-debug

now eglibc-mtrace gets included in an sdk image and not in a non-sdk image.

This does not affect builds with uclibc.

This fixes bug: [YOCTO# 2374]

(From OE-Core rev: a2ad2870a54108027ed33e977470f5f72e0bce1b)

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rpm 5.4.0: respect to the arch when choose the alternatives

There is a bug if we:
1) bitbake diffutils with MACHINE=crownbay
2) bitbake diffutils with MACHINE=qemux86
3) bitbake core-image-sato with MACHINE=crownbay

Then the diffutils.i586 would be installed to the crownbay's image, this
is because diffutils.i586 is newer than diffutils.core2, and rpm doesn't
respect to the arch priorities:

We have put the archs in order in _solve_dbpath:

crownbay/solvedb:core2/solvedb:i586/solvedb:all/solvedb

Fix rpm to respect to the order, for example, if it finds a pkg in both
core2/ and i586/, and the core2/ comes first, it should not use the one
in i586/ even if it's build time is newer.

Note: Don't worry about the _free(*ptr), it can check whether ptr is
NULL or not.

This is for the denzil branch, and the master branch also needs it.

[YOCTO #2360]

(From OE-Core rev: d76067ae100623dd6dd4858568b98522d1daf474)

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

package_rpm.bbclass: Fix incremental rpm image generation

Fix the incremental rpm image generation, it didn't work since the code
has been changed.

The btmanifest should have a ".manifest" suffix, so that it can be moved
to ${T} by rootfs_rpm.bbclass:
mv ${IMAGE_ROOTFS}/install/*.manifest ${T}/

Note: The locale pkgs would always be re-installed.

[YOCTO #2440]

(From OE-Core rev: 5149630746626c6d416f26ab9dd1c7213fcd8c50)

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

classes/base: add checksum file list at parse time

Connect the new fetcher file checksum code so that we get a list of the
files to be checksummed at parse time.

The file-checksums flag will not be read unless we are using a version
of BitBake that supports the function we call within it, so it is safe
to include this change even when the metadata will still be used with
older versions of BitBake.

Implements [YOCTO #2044].

(From OE-Core rev: f26065629b6397d129db930268b72164f8e5d3e4)

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

external-sourcery-toolchain: ignore GNU_HASH issues with its packages

(From OE-Core rev: d29852ba60e7c7657ed3a4d8ac9dacfb0de416ad)

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

Rename 'external-csl' to 'external-sourcery'

This is a rename per the purchase of CodeSourcery by Mentor Graphics
Corporation, and associated naming change.

(From OE-Core rev: dead1540d769fc91a5bd171070a5c96a9f00a2c7)

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

tcmode-external-csl: preferred external-csl-toolchain's gdbserver

(From OE-Core rev: 98df23d4673f155cc807fa7c52537b299dfd5e7a)

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

gcc: Grok for callconvention-hard to enable hard float

If callconvention-hard is set then we build gcc defaulting
to hard-float ABI

(From OE-Core rev: 88dee5cf5265d3ea26fc9d471fc6155e48935dc7)

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

eglibc: Add ARM hf dynamic linker support

The work is done in glibc upstream we backport
the relevant patches

(From OE-Core rev: 8df03a0b8bd92cf7b73b816dacf3e362cfce557a)

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

gcc-4.7: Add knowledge about arm hf dynamic loader

(From OE-Core rev: a0d9e602a778feab2201a0643fb7d555b71017b9)

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

Hob: Adjust the progress bar and set 100% only when all is done.

After parsing recipes, Hob will populate recipes and packages, which is probably
time exhaused. So, this patch is to adjust the progress bar and ensure 100% is
set if and only if all populations are done.

The patch also fixes "weird 18 second delay when parsing recipes" on build appliance.
Because Hob is doing something, but the progress bar shows 100% and wait there.

[Yocto #2341]

(Bitbake rev: 2c4a21dc8a588c8cf05549ddd9734731a46bea10)

Signed-off-by: Shane Wang <shane.wang@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

build.py: Add support for log and run filename changes

The format of the log file and run file are now selectable
using BB_LOGFMT and BB_RUNFMT, respectively.

The following values may be used:

{task} - task name
{taskfunc} - task.func or func, if task==func
{func} - function name, only available in BB_RUNFMT
{pid} - pid

The log/run files may be placed into a subdirectory that
is relative to T.

Default BB_LOGFMT is: log.{task}.{pid}
Default BB_RUNFMT is: run.{func}.{pid}

(Bitbake rev: 588da606eb81c52cb92d29041e1c67897427bfdf)

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

build.py: Add additional debug messages

We now add a debug message when entering and exiting a python or
shell function.  This makes it easier to inspect a log and figure
out the run order from the logs.

(Bitbake rev: a0f554d0e722b6705844c6031fdcafa5d1a1c8a7)

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

build.py: Cleanup exec_func_shell

exec_func_python and exec_func_shell are similar, but variable
usage has diverged sync the two up.  Since exec_func_python is first
use that as the guide for the later exec_func_shell variable naming.

(Bitbake rev: ccfe1a3a2419172799957676107f240badf3f062)

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

build.py: Add a log to capture task execution order

The new log.task_order contains an ordered list of the tasks as they
were executed in any given recipe.  The format of the lines is
<task> <pid>: <log file>

(Bitbake rev: 8662b43dcbd6e38a5b2ab9d6529af1fb08c146bf)

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

replace os.popen with subprocess.Popen

Replace os.popen with subprocess.Popen since the older function would
fail (more or less) silently if the executed program cannot be found

There is a bb.process.run() which will invoke the Popen to run command,
use it for simplify the code.

For the:
p4file = os.popen("%s%s files %s" % (p4cmd, p4opt, depot))
...
for file in p4file:
    list = file.split()

in bitbake/lib/bb/fetch2/perforce.py, it should be an error in the past,
since it didn't use readline() to read the pipe, but directly used the
split() for the pipe. Use the bb.process.run would fix the problem since
bb.process.run will return strings.

More info:
http://docs.python.org/library/subprocess.html#subprocess-replacements

[YOCTO #2075]

(Bitbake rev: 8d6700255a6d4dda403c89b171a6d4a1883e5aae)

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

replace os.system with subprocess.call

Replace os.system with subprocess.call since the older function would
fail (more or less) silently if the executed program cannot be found

More info:
http://docs.python.org/library/subprocess.html#subprocess-replacements

[YOCTO #2075]

(Bitbake rev: f5b3bf115dc1ffbfb241a49cec0fc3654cb71021)

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

bitbake: support $BITBAKE_UI environment

Patch adds support for a $BITBAKE_UI environment variable which allows
to configure the preferred user interface. Although an '-u' option
(which will override the environment variable) exists already, it was
required to specify this option on every invocation of bitbake.

Because user interface is instanciated very early in the program it is
not possible to use bitbake.conf for setting up a default.  An
environment variable (which acts in a similar category like $PAGER or
$EDITOR) is a simple way for configuring the default.

(Bitbake rev: e3c213015953d1a0afb5ef4be59e1264990e5cee)

Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: implement checksums for local files in SRC_URI

Gathers a list of paths to have checksums calculated at parse time, and
processes these when calculating task hashes. Checksums are cached with
the file's current mtime. Thus, changing any local file in SRC_URI will
now cause the do_fetch taskhash to change, thus forcing a rebuild.

This change adds very roughly about an 8% increase in parse time (a few
seconds) and maybe a few seconds during runqueue generation, so a fairly
moderate performance hit.

Note that since paths are resolved at parse time, this will not force
a rebuild when files are introduced which would cause that resolved path
to be different - for example, where a machine-specific version of a file
was added without otherwise changing the recipe. This will need to be
handled in a future update.

Code to hook this into the signature generator was courtesy of
Richard Purdie <richard.purdie@linuxfoundation.org>.

Implements [YOCTO #2044].

(Bitbake rev: c993b7c457f8b7776e8a5dff253bfa0724bc2cae)

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

bitbake: refactor out codeparser cache into a separate class

We want to be able to reuse most this functionality for the file
checksum cache.

(Bitbake rev: 0fe3cb1438d297f90dd0fc6b26362ecbff75c76d)

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

Hob: fixed the issue about the 'run image' page should have the primary bug

issues: when build a image for atom-pc, the output does not include live
image, and 'Save as template' should become the primary action on the screen

[YOCTO 2326]

(Bitbake rev: e9516c7e14d782b943cc6e6a2e5e2111edf03d8c)

Signed-off-by: Liming An <limingx.l.an@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Hob: fixed the line width of base image description view

As ui design, adjusted the line width for each base image description

[YOCTO 2310]

(Bitbake rev: 8bce98b23bc9cbc4da4464bfbaeb4e1a1aaca5c5)

Signed-off-by: Liming An <limingx.l.an@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Hob: fixed the issue about bug 'running image' workflow and related view

[YOCTO #2155]

(Bitbake rev: 5e14cb41e8ba26a3114ea2c8b8f83b271bc44b54)

Signed-off-by: Liming An <limingx.l.an@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Hob: change the build failure scenario as ui design

change the top bar display in build 'issue' page

[YOCTO #2183]

(Bitbake rev: 0705d3db1ce6d0f29301e2428c990ab0d9b2860e)

Signed-off-by: Liming An <limingx.l.an@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Hob: add '--select a machine--' and '--select a base image--' to GUI

[YOCTO #2175]

(Bitbake rev: 2729729012f035043fedc5098be2ec12b761166d)

Signed-off-by: Liming An <limingx.l.an@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

fetch2: quote/unquote url paths

This ensures we can handle things like %-encoded characters in the path
portion of urls.

(Bitbake rev: b1dbc24ebcc4e5100c32568c2c41fd982fb4bcce)

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

Hob: Get image name internally when updating the image description

The image name shouldn't be passed from outside caller, since the image
name may not in the combobox list. Getting the name from
update_image_desc() internally.

(Bitbake rev: 6635cd7bc2a448d1324e9344100f97613f4272b5)

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

Hob: Add filter for images listed in image combo

Define BBUI_IMAGE_WHITE_PATTERN variable to indicate which image is
allowed to be displayed in image combobox.

Define BBUI_IMAGE_BLACK_PATTERN variable to indicate which image is NOT
allowed to be displayed in image combobox.

This fixes [YOCTO #1581]

(Bitbake rev: 960461bc12945675af3081eb469f932f4a6eb1cd)

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

Hob: reimplement the proxy page

This patch is to reimplement the proxy page in the "Advanced Settings" dialog
per the new design in https://bugzilla.yoctoproject.org/attachment.cgi?id=442
and https://bugzilla.yoctoproject.org/attachment.cgi?id=443.

[Yocto #2247]

(Bitbake rev: 911a60c09c1539a3f10c2bcdb26d40e458c31303)

Signed-off-by: Shane Wang <shane.wang@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

fetch2: Also exclude urls containing wildcards from checksumming for now

Without this change, bitbake will try and checksum a wildcard which
will lead to fetch failures.

(Bitbake rev: ac53b88be58b0bed21730c0b61a8fc8e801a2f1b)

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

tune-mips64.inc: Add new tune file for mips64 big-endian

(From OE-Core rev: e6333825c3482a559a0c0499e17f8f48d3042ddf)

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

xserver-xorg: Fix build for mips64

mips64 N64 ABI has different size for int and pointer
which means the compiler will complain about conversions
which works fine for 32 bit since sizes are same.

amd64 is taken care of already so we take care of mips64
with this patch by understanding the difference in sizes
in compiler.h

(From OE-Core rev: de21917ba83f104a2d53941f40a9b2b8d35dce49)

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

site/sh-common: Add missing caches variables to build glib-2.32

This is on the same tone as done for other architectures

(From OE-Core rev: 7bdc3a07d3212d7fc820c77ba509e476cc7aa672)

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

netbase: Add interface files for qemumips64 and qemush4

These files are alike other qemu machines they
help in creating a working network interface when
using qemumips64 or qemush4

(From OE-Core rev: 1e7406e4f2f93b6b4fa75716c4e901a96206be69)

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

runqemu: Add qemush4 and qemumips64 knowledge

New machines need to be added and they also
have different kernel commandlines

(From OE-Core rev: 3a5432aec0faea49d2c04984cd169ceb35bba89f)

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

libc-package: Add sh4 and mips64 to arch options

needed for new architecture support

(From OE-Core rev: 75167eb890d8e4683f1e3852e3b77c67d160bea7)

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

eglibc-2.15: Support mips64

Add mips64 to binary locale arches
Enabling ports in addons causes wrong
machtype to be chosen for mips64 and
it always chose mips32 which is wrong
for mips64 so enable all the addons
we have and we were enabling them anyway

(From OE-Core rev: da7f9093ad7aa4188cc3d420668b0bda7d950a0f)

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

kernel-arch.bbclass: Map mips64{el} to mips KARCH

(From OE-Core rev: 22ce249a814f750bc47ead9ce281fcbd63c775c8)

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

binutils: Default to n64 when configured for mips64

Default ABI for mips64 is n64 so make it default in binutils

(From OE-Core rev: 99f859612e16d636702f9bb11825259d90e468f2)

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

gcc-4.6, gcc-4.7: Add support for building mips64 cross compiler

Defaults to n64 ABI

(From OE-Core rev: 253f1f3aebd1a90c55efcc6d1f585204e7f3c371)

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

site: Add mips64 eglibc and uclibc site files

(From OE-Core rev: 2e315d2d0de215811e8caa80b1f13fd0e9fa53e8)

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

insane.bbclass: Add mips64{el} to known machines

(From OE-Core rev: 77dc886107e1b9cdaf6803141dc97c3ebfc22910)

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

bb/fetch2/__init__.py: Don't try to compute checksums for directories

In this way we avoid failing the build while trying to fetch local
directories.

[YOCTO #2475]

(Bitbake rev: 39adb5741d9eee0879d3181be505400dffc60804)

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

hob: update required pygtk to 2.22.0 and gtk+ to 2.20.0

(Bitbake rev: e67ccbbeaecd2d9f3267995c590052c6a3c5b9ce)

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

lib/bb/ui/crumbs/[package|recipe]selectionpage: include button

The indication of the included packages/recipes in the top right of the
GUI, per design, is a button which switches the Notebook to the included
tab.

(Bitbake rev: 494534cae6211d9055098562df1f35b56858f3ec)

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

lib/bb/ui/crumbs/imageconfiguration: fix combo boxes drop-down width

The width of the combo boxes drop down should match the combo box
itself.

(Bitbake rev: 93e66f494730d0840dd1d4f3ed924d91e6489995)

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

lib/bb/ui/crumbs/hobwidget: use light text in HobIndicator

Dark text on a dark background is difficult to read. Further the design
document uses white text.

(Bitbake rev: a84304ccf06a012817a5e6495d53147efcbed960)

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

lib/bb/ui/crumbs/: replace custom HobNotebook

The custom HobNotebook widget was implemented to address perceived
deficiencies in the gtk.Notebook API.

Recent inspection reveals that the API is capable of all that Hob
requires of it and therefore maintaining a custom class to provide
similar functionality does not make sense.

Addresses [YOCTO #2276]

(Bitbake rev: e683caa9863bbb52480346669806f22173629a5e)

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

Hob:add function about clicked a row to load image file for 'my images' page

[YOCTO #2238]

(Bitbake rev: b72ee3cf29a5badca3f8c1342356164270615aad)

Signed-off-by: Liming An <limingx.l.an@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Hob: Apply a line length to consistantly display image descriptions

[YOCTO #2310]

(Bitbake rev: e9d65e06aaf20bd9f5ca02544c1bf976c6bff9be)

Signed-off-by: Liming An <limingx.l.an@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

knotty2: return exit code

'bitbake -u knotty2' never failed which was caused by main() not
propagating the exit code.

(Bitbake rev: 243e4ba405b88fd466bec7b4aa042d20aa6121ff)

Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake/fetch: Spell out which fetcher backends support and recommend checksums

There were some hardcoded behaviours in the system for which backends
support checksums verses which backends recommend them verses which
don't recommend them.

This moves the functionality into specific fetchers and then makes the
general code generic. This cleans up the codebase and fixes some corner
cases such as trying to checksum directories returned by the git fetcher.

(Bitbake rev: ef6d268f7b8527541a7fb044cf95a973be4097f4)

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

bitbake/fetch: If checksums are available, check them for any file, local or remote

Currently, checksums are only checked for remote files. This changes
the check to apply to any file we have checksum data for. A mismatch
against a file is fatal but in the local case, no warnings are shown
about missing checksums.

(Bitbake rev: 80411b3acf2173a4a7d415102d16676eb98363e3)

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

bitbake/fetch2: Improve visibility of checksum warnings when fetching from mirrors

When fetching from mirrors, checksum errors would get buried in the
logs. This raises their profile so a warning is logged on the console
when fetcher checksum issues are encountered, even if other attempts
are made to get the file (which may or may not have the same issue).

(Bitbake rev: d43fafd7f01b5534499b45213197d8ccececdbc4)

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

bitbake/fetch2: Improve stamp creation in mirror code

Under certain contitions such as local file:// mirrors of all files,
bitbake might not check checksums. This is not desirable and we should
always check the checksum where possible.

This only exception to this is the git mirror tarball case where the
checksum is unknown but we need to create the .done stamp for other
reasons. This patch preserves thta functionality but in a more specific
section of code where this doesn't interfere in normal checksums.

(Bitbake rev: 2592a43e422c06a6f6174ada562e9c8ba25e1ec1)

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

bitbake/fetch2: Remove hardcoded file removal and use clean method

Under certain circumstances, bitbake could remove files which were
outside its control since it unconditionally removes ud.localpath.
In the file:// case this can point at external paths.

This patch converts it to use the clean() method which will remove
files when its safe to do so but not in the file:// case since the
file:// handler has an empty clean method. This means bitbake no
longer removes files outside its control and is generally much safer.

(Bitbake rev: 9ad1cebe2220b41da2141cdb3f0a403abb46d77c)

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

sstate.bbclass: Make sure we don't have an empty fixmepath file

Jason Wessel noticed that a package without any fixmepath entries would
generate a sed warning about no input files.  This patch resolves that
by ensuring that an empty fixmepath file never gets written into the
sstate archive.  Also we avoid a second message by only doing xargs if
we got input.

(From OE-Core rev: 326563d5a897ae2dba7cfd8d73579d3d979d72c8)

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