gvfs: switch smb support to samba4

samba3 is deprecated upstream.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

kodi: switch smb support to samba4

samba3 is deprecated upstream.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

mpd: switch smb support to samba4

samba3 is deprecated upstream.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

icu: bump to version 55.1

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/download: quick fix to avoid breaking on custom toolchains

When the user selects a custom toolchain to be downloaded, there's no
hash for that toolchain, so the download fails, now that hashes are
mandatory.

Fix that by simply exiting as if there was no error, until we have a
better fix...

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Makefile: remove the graphs/ dir on 'make clean'

Currently, a 'make clean' leaves the graphs/ subdirectory in the
output directory. This commit defines a GRAPHS_DIR variable, used by
the different graph-generating targets, and which gets cleaned up in
the 'clean' target.

[Thomas: use the new GRAPHS_DIR variable in more places, as suggested
by Yann.]

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/download: warn when there's no .hash file

Instead of silently accepting a missing .hash file, print a warning.

This can be grepped from a build log, to find packages that still have
no hash, with the long-term goal of adding hashes for all packages.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/download: always fail when there's no hash

At the time we introduced hashes, we did not want to be too harsh in the
beginning, and give people some time to adapt and accept the hashes. So
we so far only whined^Wwarned about a missing hash (when the .hash file
exists).

Some time has passed now, and people are still missing updating hashes
when bumping packages.

Let's make that warning a little bit more annoying...

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/download: properly catch missing hashes

When checking hashes reports no hash for a file, and this is treated as
an error (now: because BR2_ENFORCE_CHECK_HASH is set; later: because
that will be the new and only behaviour), exit promptly in error.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/download: return different exit codes for different failures

Return different exit codes depending on the error that occured:

  0: no error (hash file missing, or all hashes match)
  1: unknown option
  2: hash file exists, but at least one hash in error
  3: hash file exists, but no hash for file to check
  4: hash file exists, but at least one hash type unknown

This will be used in a later patch to decide whether the downloaded file
should be kept or removed.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Samuel Martin <s.martin49@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/download: add explicit no-hash support

Add support to explicitly state that an archive has no hash.

This can be used for archives downloaded from a repository, like a
git-clone or a subversion checkout, or using the github helper.

This will come in handy when we'll eventually make hashes mandatory as
soon as a .hash file exists: for some packages, like gcc, some versions
are downloaded as archives from upstream, while other versions may come
from a GitHub repository (via the github herlper).

In this case, a .hash file would exist, that contains hashes for the
downloaded tarballs, but archives downloaded from the repository would
not have a hash (since it is currently not possible to have reproducible
such archives). So, we'd need a way to explicitly state there is no
hash, on purpose, for those archives.

So, add 'none' as a new type of hash.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package infra: do not check hashes when downloading from a repository

When downloading from a repository, we have no way to ensure the
reproducibility of the generated archives, so we can't check the hashes.

Do not specifiy a hash file in those cases.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/download: make hash file optional

Currently, specifying a hash file for our download wrapper is mandatory.

However, when we download a git, svn, bzr, hg or cvs tree, there's by
design no hash to check the download against.

Since we're going to have hash checking mandatory when a hash file
exists, this would break those downloads from a repository.

So, make specifying a hash file optional when calling our download
wrapper and bail out early from the check-hash script if no hash file is
specified.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/softether: bump version

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

softether: add patch to fix build on uClibc toolchain without ifaddr support

This commit adds a patch to softether to make sure it doesn't use the
ifaddrs functions when they are not provided by the C library, which
can be the case with certain external uClibc toolchains.

Fixes:

  http://autobuild.buildroot.net/results/199/19992951da34a703a476dc62ea92e12927c8a02d/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/ti-gfx: bump version, add hashes

We don't bump the softfp version, because there's still no newer release
(probably softfp will never be updated).

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/mc: Fix xlib_libX11 dependency

On Peter´s autobuilder building mc fails:
http://autobuild.buildroot.net/?reason=mc-4.8.14

The reason can be found in
http://autobuild.buildroot.net/results/e25/e25587c17c2e8ddae83f9ed02541931f510bfcd6//build-end.log

checking for X... libraries /usr/lib, headers /usr/include

where configure picks up xorg libraries although no x11 stuff is enabled in the
defconfig used for this build.

This leads to subsequent errors visible in
http://autobuild.buildroot.net/results/e25/e25587c17c2e8ddae83f9ed02541931f510bfcd6//mc-4.8.14/config.log

mips64el-ctng_n64-linux-gnu-gcc: ERROR: unsafe header/library path used in cross-compilation: '/usr/lib'

causing most configure checks to fail afterwards. To fix the problem the mc
package should honour an optional dependency to the xlib_libX11 package.

Fixes
http://autobuild.buildroot.net/results/e25/e25587c17c2e8ddae83f9ed02541931f510bfcd6/
http://autobuild.buildroot.net/results/b7a/b7a2b69c53e91e0173b7f0c241d7e06b4382c52a/
http://autobuild.buildroot.net/results/805/805675236190f7f5f42d498f4a9f9bf0e1bab648/
and others

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Reviewed-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

docs: add links to the rest of the licenses

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

linphone: install to staging

Make sure the liblinphone library gets installed too.

Signed-off-by: Sagaert Johan <sagaert.johan@proximus.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

qt5: update hashes for 5.4.1

When updating Qt to 5.4.1 the hash files were omitted

Signed-off-by: Will Wagner <will_wagner@carallon.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mpg123: bump to version 1.22.1

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/linux-firmware: install AMD Radeon firmware

Signed-off-by: David BARBION <david.barbion@ext.leroymerlin.fr>
[yann.morin.1998@free.fr:
  - simplify by copying the full directory
  - fix title
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/linux-firmware: add support to copy entire directories

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: David BARBION <david.barbion@ext.leroymerlin.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

imagemagick: bump to version 6.9.1-0

Also update hash file.

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/powertop: depends on c++

Fixes configure error due to missing c++

checking for library containing pthread_create... no
configure: error: libpthread is required but was not found

Quote from config.log:

configure:21552: checking for library containing pthread_create
configure:21583: false -o conftest    -Os    conftest.cpp -lintl  >&5
[...]
ac_cv_env_CXX_set=set
ac_cv_env_CXX_value=false

using this defconfig

BR2_TOOLCHAIN_BUILDROOT_WCHAR=y
BR2_PACKAGE_POWERTOP=y

Please note the number of .cpp files belonging to powertop

output/build/powertop-2.7$ find -iname *.cpp | wc -l
59

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/powertop: handle ncursesw properly and depend on wchar

The powertop code uses some wide-char functions, so the package should
depend on wide-char support. Also, when wide-char support is
available, the wide-char variant of ncurses may be built, but powertop
detection logic isn't quite up to the game, so we help him a little
bit.

Thanks to Gustavo Zacarias for doing some additional investigation and
proposing the solution implemented in this patch.

Fixes
http://autobuild.buildroot.net/results/913/913cea22f8a8f5902d8da5f64c3fce056d66790f/
http://autobuild.buildroot.net/results/6e1/6e11fa2a7405a69c59ced046b92ff08660c4aab7/
http://autobuild.buildroot.net/results/1d3/1d3323b2afaefa7989854dbccf92015731199e66/
http://autobuild.buildroot.net/results/1e3/1e31d412d8b3a38a375ad0be8f696bee993ec297/
http://autobuild.buildroot.net/results/2ff/2ff511eb8d00b94aca68427446e2d0f6e4317a5a/
and maybe others

[Thomas:
 - do not force select NCURSES_WCHAR, instead improve the detection
   logic to properly use it when available. Solution provided by
   Gustavo.
 - improve the commit log.]

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gnuradio: depends on shared libs

volk must be compiled with toolchain built with shared lib to avoid :
attempted static link of dynamic object `../lib/libvolk.so.0.0.0'

[Thomas:
 - rebase after the non-largefile removal.
 - comment must be "dynamic library" not "shared library".]

Signed-off-by: Gwenhael Goavec-Merou <gwenhael.goavec-merou@trabucayre.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

jamvm: add depends on toolchain w/threads

Signed-off-by: Matt Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs: add Freescale SABRE Auto board support

SABRE Board for Automotive Infotainment (SABRE Auto, a.k.a. SABRE-AI) is
Freescale's evaluation board based on the i.MX 6 ARM Cortex-A9 applications
processor.

Those defconfigs are an adaptation of freescale_imx6{q,dl}sabresd_defconfig for
SABRE Auto, and are thus based on Freescale "official" git repo on
git.freescale.com and SW release 3.10.17_1.0.0_ga.

Signed-off-by: Vincent Stehlé <vincent.stehle@freescale.com>
Cc: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Gilles Talis <gilles.talis@gmail.com>
Cc: Peter Korsgaard <jacmet@sunsite.dk>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Gary Bisson <bisson.gary@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

imx6sabresd: rename to imx6sabre

Rename imx6sabresd board folder to imx6sabre, to prepare for Sabre Auto
addition. Update doc, link and defconfigs accordingly.

Signed-off-by: Vincent Stehlé <vincent.stehle@freescale.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Eric Bénard <eric@eukrea.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Gilles Talis <gilles.talis@gmail.com>
Cc: Gary Bisson <bisson.gary@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/nvidia-driver: add option to install CUDA MPS server

Using CUDA with NVidia requires those two programs if one wants to use
more than one program doing CUDA at the same time.

This is only available on x86_64.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Al West <al.west@v-nova.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/nvidia-driver: CUDA support does not require X.org

Currently, nvidia-driver always installs the X.org driver, assuming this
is a requirement for all functionalities provided by nvidia-driver, thus
pulling in quite a bunch of X.org packages.

However, it is perfectly legit to be doing CUDA (and OpenCL) work
without the full X.org stack, and indeed the NVidia CUDA and OpenCL
libraries do not require the X.org stack.

Split the configuration so that it is posible to install the different
parts independently from each others, so that CUDA can be installed all
on its own.

Reported-by: Al West <al.west@v-nova.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Al West <al.west@v-nova.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

liberation: bump to version 2.00.1

Also add hash file and switch license.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package: new category for fonts, icons, sounds and themes

Move all of the scattered fonts, icons, sound packs and themes into a
new category for them.
In preparation for the inclusion of more fonts.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

jimtcl: fix installation in BR2_STATIC_LIBS case

As noticed by Yann E. Morin in the review of
http://patchwork.ozlabs.org/patch/429533/, there was something fishy
in the jimtcl installation logic:

  ln -s libjim.$(JIMTCL_LIB) $(STAGING_DIR)/usr/lib/libjim.so

where JIMTCL_LIB has the value 'a' for BR2_STATIC_LIBS=y builds. Which
means we're linking libjim.so to libjim.a. Not great.

This commit therefore reworks the installation logic of the jimtcl.mk
package to install the shared library when BR2_STATIC_LIBS is not set,
and the static library when BR2_STATIC_LIBS is enabled. The macro
JIMTCL_INSTALL_LIB now takes as argument where the library should be
installed, so that it can be used for both the target and staging
installations.

Note that we can only either build the shared library *or* the static
library with the jimtcl build system. There is no possibility of
building both.

Reported-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

jimtcl: change the way of defining the JIMTCL_INSTALL_LIB variable

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain-common: drop BR2_LARGEFILE

It's now unused so remove it.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/uclibc: drop BR2_TOOLCHAIN_BUILDROOT_LARGEFILE

Kill the option to build the non-largefile variant and remove the hidden
option since it's now unused.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

packages: remove non-lfs enabler patches

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

docs/manual: remove references to largefile

Switch from the ACL package example to RRDTOOL since ACL has no dual
(depends + select) dependencies any more and doesn't fit the example
scenario.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

packages: remove (non-)lfs dependencies and tweaks

Now that largefile is mandatory removes package dependencies and
conditionals.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

boot/gummiboot: drop largefile depends

Drop the buildroot toolchain non-largefile support option since it's
being removed.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

system: drop largefile depends

Drop the buildroot toolchain non-largefile support option since it's
being removed.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

busybox: force lfs build

Drop the buildroot toolchain non-largefile support option since it's
being removed.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs: drop largefile option

Drop the buildroot toolchain non-largefile support option since it's
being removed.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package infra: drop non-lfs support

Now that largefile is mandatory remove support for non-lfs
tweaks/variables in the package infra and the gcc build.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain: remove LARGEFILE selects

There's no need for toolchains or the user to declare largefile support
since it's now mandatory.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain-common: default to Y for LARGEFILE

This will allow us to remove largefile handling in the tree without
breaking things while doing so.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/uclibc: default to Y for LARGEFILE

This will allow us to remove largefile handling in the tree without
breaking things while doing so.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain/helpers: make LFS mandatory for uclibc toolchains

As discussed on the mailing list drop the non-largefile option for
toolchains.
The size delta is minimal and it just complicates package dependencies.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain/helpers: add check for mandatory uClibc options

We currently only check that the Buildroot configuration matches what is
available in the toolchain.

Since we're going to remove the check for LFS and make it a mandatory
feature, we will lose the corresponding buildroot option, so we won't be
able to use check_uclibc_feature as-is.

Introduce a magic value passed as the buildroot option name to recognise
checks for mandatory uclibc options that do not have a corresponding
option in buildroot.

If the buildroot option name is empty then the check is against a
mandatory uclibc option.

If a mandatory uclibc option is missing we reject the toolchain as being
unusable by buildroot.

[Thomas: minor tweaks in comment, remove space instead of tab.]

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/gtk2-themes: remove

The category just contains gtk2-theme-hicolor... which is a duplicate of
hicolor-icon-theme, d'oh!
Remove it and select the only one via legacy.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

bmon: fix license information

bmon is licensed under two licenses BSD-2c and MIT.

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

matchbox: adjust style

Remove stray comments, adjust indentation, kill empty defines.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

matchbox-wm: add hash file

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

matchbox-startup-monitor: add hash file

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

matchbox-panel: add hash file

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

matchbox-lib: add hash file

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

matchbox-keyboard: add hash file

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

matchbox-fakekey: add hash file

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

matchbox-desktop: add hash file

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

matchbox-common: add hash file

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

matchbox: switch homepage and enhance description

The old homepage is dead so switch to the new URL.
Also make the description somewhat better.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

mongoose: add hash file

Signed-off-by: Davide Viti <zinosat@tiscali.it>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

atk: bump to version 2.16

Also update hash file.

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

cairo: bump to version 1.14.2

Add hash file and remove upstream patches.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

vala: bump to version 0.26.2

Also add hash file.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/graph-depends: allow excluding virtual packages

Like for --stop-on, make --exclude recognise the keyword 'virtual',
to stop on virtual packages (as explained in the help...).

Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/x11r7/xserver_xorg-server: Glamor depends on dri3

Fixes

../../../glamor/glamor_egl.c:603:25: error: unknown type name 'RRProviderPtr'
                         RRProviderPtr provider,
                         ^
../../../glamor/glamor_egl.c:650:1: error: unknown type name 'dri3_screen_info_rec'
 static dri3_screen_info_rec glamor_dri3_info = {
 ^

using this defconfig

BR2_TOOLCHAIN_BUILDROOT_LARGEFILE=y
BR2_TOOLCHAIN_BUILDROOT_WCHAR=y
BR2_TOOLCHAIN_BUILDROOT_CXX=y
BR2_INIT_NONE=y
BR2_ROOTFS_DEVICE_CREATION_DYNAMIC_EUDEV=y
BR2_PACKAGE_MESA3D=y
BR2_PACKAGE_MESA3D_GALLIUM_DRIVER_NOUVEAU=y
BR2_PACKAGE_MESA3D_GALLIUM_DRIVER_R600=y
BR2_PACKAGE_MESA3D_GALLIUM_DRIVER_SVGA=y
BR2_PACKAGE_MESA3D_GALLIUM_DRIVER_SWRAST=y
BR2_PACKAGE_MESA3D_DRI_DRIVER_SWRAST=y
BR2_PACKAGE_MESA3D_DRI_DRIVER_I915=y
BR2_PACKAGE_MESA3D_DRI_DRIVER_I965=y
BR2_PACKAGE_MESA3D_DRI_DRIVER_RADEON=y
BR2_PACKAGE_MESA3D_OPENGL_EGL=y
BR2_PACKAGE_MESA3D_OPENGL_ES=y
BR2_PACKAGE_XORG7=y
BR2_PACKAGE_XSERVER_XORG_SERVER=y
BR2_PACKAGE_XSERVER_XORG_SERVER_MODULAR=y
BR2_PACKAGE_XDRIVER_XF86_VIDEO_INTEL=y

RRProviderPtr is defined in dri3/dri3.h, so make sure glamor is enabled only
if dri3 is enabled, too.

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-ws4py: new package

[Thomas: fix <pkg>_LICENSE_FILES, since there is no file named
'LICENSE' in the tarball.]

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

docs/manual: document new graph-depends options

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Francois Perrad <fperrad@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/graph-depends: add option to completely exclude a package

Similar to --stop-on, but also omits the package from the graph.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Francois Perrad <fperrad@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/graph-depends: accepts globs to stop on package

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Francois Perrad <fperrad@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/graph-depends: add option to stop on specific packages

Add a new option to graph-depends, that users can set to stop the graph
on a specific (set of) package(s).

This accepts any actual package name, or the 'virtual' keyword to stop
on virtual packages.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Francois Perrad <fperrad@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/libseccomp: needs headers >= 3.12

EM_AARCH64 was added to linux kernel with version 3.12:
https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/log/include/uapi/linux/elf-em.h?h=linux-3.12.y

Fixes
http://autobuild.buildroot.net/results/4cd/4cd6aaccde9fb0f80e49133b477f330b601f4a63/
http://autobuild.buildroot.net/results/bd1/bd105e821de583e582199aae31fab13ef17e67a0/
http://autobuild.buildroot.net/results/092/092143df2b91d50850067274da531c713f2ccb72/
http://autobuild.buildroot.net/results/814/8149cf0c3ef3ef8d91b5891591a71691dc3d7427/

[Thomas: fix Config.in comment.]

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

glib-networking: bump to version 2.42.1

Add hash file and drop libgcrypt tweak since it's no longer required.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/x11r7/xserver_xorg-server: add --enable-libdrm for udev

If the libdrm package is enabled, the system is built with an udev
provider and MESA3D_DRI_DRIVER is disabled this adds a --disable-libdrm
to configure options, but udev still requires libdrm and results in a
build failure since libdrm include paths aren't filled in (missing drm.h
include error).
So add --enable-libdrm for the udev scenario to counteract the previous
--disable-libdrm.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

batctl: fix musl builds

Fixes
http://autobuild.buildroot.net/results/0701b880d0f82ef92d742346bfa97bb2af216ed3/

Signed-off-by: Alex Suykov <alex.suykov@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

ccache: fix the indentation

Commit "packages: indentation cleanup" broke the build of master.
This commit reverse the patch on just ccache package and fix that issue.

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libgtk3: bump to version 3.14.11

Remove upstream patch.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

harfbuzz: bump to version 0.9.40

Also add hash file.
And do an autoreconf run since otherwise the libtool patch fails.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

hicolor-icon-theme: bump to version 0.15

Also add hash file.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

polarssl: switch download URL

Old wget versions aren't very happy with https moves/alt names hence
complain when trying to download from said sites.
Since polarssl is now mbed tls and everything got renamed switch to the
new URL to avoid this.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

packages: indentation cleanup

This commit doesn't touch infra packages.

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

jsoncpp: fix build for non C++11 toolchains

Add a patch removing reference to std::snprintf that is only in C++11.

Fixes:
http://autobuild.buildroot.net/results/b39/b390beeb9a6fea12f1a4d9f2ff9a8a483628cad2/
http://autobuild.buildroot.net/results/d18/d18b3e116b680a3897fbea296938e900a1e5e80a/
http://autobuild.buildroot.net/results/1b2/1b20a2f81dd2b409785930705222bc03576c4907/

and many more.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

mesa3d: specify llvm-prefix

When it's not defined an automatic search is triggered, which can find
/usr/bin/llvm-config on distributions that include all development files
(like gentoo) and have desktop packages installed.
So point it towards the staging dir which is where it should live
eventually if/when we've got it to avoid header/include path polution
that otherwise leads to build failures.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

pango: adjust indentation

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

pango: bump to version 1.36.8

Also add hash file.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

pixman: add hash file

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

perl: use <pkg>_EXTRA_DOWNLOADS

Instead of manually using the DOWNLOAD macro (which should remain an
internal macro), this commit converts the Perl package to use
<pkg>_EXTRA_DOWNLOADS, now that it has been extended to allow full
URLs.

[Thomas: as suggested by Yann, keep comment explaining how we handle
perl-cross.]

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

docs/manual: update documentation about <pkg>_SOURCE, <pkg>_PATCH and <pkg>_EXTRA_DOWNLOADS

This commit updates the Buildroot manual for the variables used to
indicate where to download the source code from:

 - It updates the description of <pkg>_SOURCE to make it clear that
   Buildroot assume that the tarball is hosted at <pkg>_SITE.

 - It updates the description of <pkg>_PATCH to indicate that not only
   file names (assumed to be hosted at <pkg>_SITE) can be used, but
   also full URLs. This allows to match with what the current code is
   doing.

 - It updates the description of <pkg>_EXTRA_DOWNLOADS to also
   indicate that full URLs are now accepted, following the change made
   in the previous commit.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

pkg-generic: allow full URLs for <pkg>_EXTRA_DOWNLOADS

The current logic for <pkg>_EXTRA_DOWNLOADS assumes that it is a list
of files, all hosted at <pkg>_SITE. However, just like for
<pkg>_PATCH, it may be useful to specify <pkg>_EXTRA_DOWNLOADS entries
that are hosted on a different site than the package <pkg>_SITE.

This commit implements this, by re-using the same logic as the one
used for <pkg>_PATCH.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

pkg-generic: refactor the "Downloading" message logic

In the .stamp_downloaded hook, the logic to decide whether or not to
display the "Downloading" message is treating the check of
<pkg>_SOURCE as a special case. But in fact, there is no real reason
to do so: the existing loop used for <pkg>_PATCH and
<pkg>_EXTRA_DOWNLOADS could work just as well.

This commit therefore refactors this piece of code, to have a single
loop checking <pkg>_SOURCE, <pkg>_PATCH and <pkg>_EXTRA_DOWNLOADS.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

pkg-generic: take into account <pkg>_EXTRA_DOWNLOADS to display "Downloading" message

The .stamp_downloaded target displays the "Downloading" message even
if there is really something to download. However, this logic only
checks for <pkg>_SOURCE and <pkg>_PATCH: it does not check if
something needs to be downloaded in <pkg>_EXTRA_DOWNLOADS.

This commit fixes that, which makes sure that the "Downloading"
message is displayed if one of the items in <pkg>_EXTRA_DOWNLOADS has
not yet been downloaded.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

pkg-generic: fix the logic showing the "Downloading" message

Even though the .stamp_downloaded target is executed for each package
being built, the pkg-generic infrastructure tries to not display the
"Downloading" message when there is in fact nothing to download.

Unfortunately, the logic was incorrect for the patch download: it
forgot the fact that <pkg>_PATCH can contain either file names (in
which case we assume the patch should be downloaded from <pkg>_SITE),
or full URLs. The latter case was not properly handled, as we were
checking if $(DL_DIR)/<full URL> existed, while we should be testing
if $(DL_DIR)/`basename <full URL>` exists.

This patch fixes that, which makes sure the "Downloading" message is
displayed only when necessary.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

linux: use the package infrastructure to download patches

The linux package has a special handling of patches, with quite a bit
of legacy in it. A problem caused by this special handling is that the
linux package calls directly the DOWNLOAD_WGET macro, which means that
the package infrastructure isn't aware of which patches get
downloaded, and it prevents doing changes inside the package download
infrastructure.

This commit changes the handling of patches in the linux package in
the following way:

 * The LINUX_PATCHES variable is kept as is: it lists all the patches
   mentioned in the Config.in option BR2_LINUX_KERNEL_PATCH. This
   option can contain http://, ftp://, https:// URLs, path to local
   files or local directories.

   This variable is *not* used by the generic package infrastructure,
   so it is purely internal to the Linux package.

 * The LINUX_PATCH variable is now filled in with the list of patches
   that should be downloaded. It is derived from LINUX_PATCHES by
   filtering the patches that have http://, ftp:// or https:// in
   their path. Since <pkg>_PATCH is handled by the package
   infrastructure, it means that those patches are now automatically
   downloaded and applied by the package infrastructure.

 * The LINUX_APPLY_PATCHES hook is renamed to
   LINUX_APPLY_LOCAL_PATCHES, because it is now only responsible of
   applying local patches: remote patches are handled by
   LINUX_PATCH. The implementation of the hook is changed to filter
   out the patches that have already taken care of by LINUX_PATCH, so
   that we only iterate through the list of local patches or local
   patch directories.

[Thomas: adjust comment in the code according to Yann comments.]

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

libgtk2: adjust style

Cleanup indentation and mention why we need AUTORECONF.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libgtk2: drop libxcomposite mandatory dep

It's not selected, not required and duplicate with the autodep.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libgtk2: bump to version 2.24.27

Also add hash file.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

logrotate: add hash file

Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau.dev@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>