lighttpd: run as non-privileged user www-data

Previously the stock lighttpd configuration was patched to run as user root,
because we do not have a lighttpd user.
Patch it to use www-data instead, which is a user available in the default
skeleton, and is the same other webserver packages (like nginx) use.

Signed-off-by: Floris Bos <bos@je-eigen-domein.nl>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libtasn1: bump to version 4.5

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

libmicrohttpd: bump to version 0.9.41

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

stunnel: bump to version 5.17

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

gst-plugins-bad: unbreak vp8 support

Fixes
http://autobuild.buildroot.net/results/84e/84e7a2df4a8de2ca1ce075a8f5f32a94b7f45ce8/
http://autobuild.buildroot.net/results/db8/db8309c8cc6d4fb8a28d17c3b1ed71c49277d212/
http://autobuild.buildroot.net/results/5e4/5e424a25ee6796207e7a85ecbc80ac4428b0522e/
http://autobuild.buildroot.net/results/3b0/3b0e0356549d40edbe1c9a86214437e0f81fb3a3/

And many more.

gst-plugins-bad didn't have any explicit libvpx handling, so it would be
built with/without support for it depending on build order.

More importantly, the recent libvpx bump dropped the legacy (prefix-less)
symbols from the header file and gst-plugins-bad's handling of the non
legacy symbols was buggy.

Fix it by adding an explicit vp8 suboption and add a patch to fix (drop) the
legacy symbol handling.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/clamav: bump version to 0.98.7, enable ipv6 support

Clamav uses AC_TRY_RUN in m4/reorganization/code_checks/ipv6.m4 to check
for ipv6 support, which is not cross-compile safe. Since buildroot
supports ipv6 out-of-the-box now this patch forces ipv6 support.

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

ncftp: add hash file

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gawk: bump to version 4.1.2

Drop LIBTOOL/GETTEXTIZE as well since it's no longer bundled with a beta
version.

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

btrfs-progs: bump to version 4.0

Switch SITE to use BR2_KERNEL_MIRROR as well.

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

libcurl: security bump to version 7.42.1

Fixes:
CVE-2013-3153 - sensitive HTTP server headers also sent to proxies.

And drop upstream patches.

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

pcre: bump to version 8.37

Security patch now upstream so drop it.
Also switch to bz2 compression.

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

linux: bump default to version 4.0.1

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

linux-headers: bump 3.{10, 14, 18, 19}.x and 4.0 series

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

dhcpcd: add upstream patch to fix build issue

Fixes http://autobuild.buildroot.net/results/523/523858fb930e74fe28f24989ff0697693b5fdd67/

Patch from http://roy.marples.name/projects/dhcpcd/vpatch?from=e05490e2d12d03b1&to=c32714545158ca83

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Revert "dhcpcd: fix ipv6 related compile failure"

This reverts commit 6b1f66643d782fc6f00e7436d41ceea6d88ad6bf.

It turns out to not really fix the build errors after all, so revert it.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libwebsockets: add LWS_ prefix to LWS_WITHOUT_TESTAPPS=ON cmake option

Currently, libwebsockets.mk is passing -DWITHOUT_TESTAPPS=ON to disable
the building of libwebsockets test applications, but this is doing nothing
because the option is really named  LWS_WITHOUT_TESTAPPS.
This commit fixes that by using the correct option name.

Signed-off-by: Baptiste Moraine <baptiste.moraine@kapelse.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

dhcpcd: fix ipv6 related compile failure

Add '-D_GNU_SOURCE' to gain missing struct in6_pktinfo definition.

Fixes [1]:

ipv6.h:211:41: error: invalid application of 'sizeof' to incomplete type 'struct in6_pktinfo'
  unsigned char sndbuf[CMSG_SPACE(sizeof(struct in6_pktinfo))];
                                         ^
ipv6.h:195:38: error: invalid application of 'sizeof' to incomplete type 'struct in6_pktinfo'
 #define IP6BUFLEN (CMSG_SPACE(sizeof(struct in6_pktinfo)) + \
                                      ^
ipv6.h:214:23: note: in expansion of macro 'IP6BUFLEN'
  unsigned char rcvbuf[IP6BUFLEN];

[1] http://autobuild.buildroot.net/results/523/523858fb930e74fe28f24989ff0697693b5fdd67

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

qt: QT_COORD_TYPE_DOUBLE should only be set on soft float

If QT_COORD_TYPE_DOUBLE is forced and hard float is used, Qt compilation
breaks because it can't find hard coded operations on double.

Moreover, QT_COORD_TYPE is float only if QT_NO_FPU is set.
(cf src/corelib/global/qglobal.h)

So, we can safely make the QT_COORD_TYPE_DOUBLE option depend on
BR2_SOFT_FLOAT.

Reported-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Richard Genoud <richard.genoud@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gdb: add hash file

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

python3: needs dynamic library support

Fixes:
http://autobuild.buildroot.net/results/c65/c65c2b2c4c2ca328c4f26ba38798c40ab02c8772/
http://autobuild.buildroot.net/results/e0d/e0db38ca12dbe664d472c4f6f3f1ade4150119c3/
http://autobuild.buildroot.net/results/8e1/8e196d05dd65cdd782c8d41d937a7d9d2ef7dc1a/
http://autobuild.buildroot.net/results/e3a/e3a71f5d05a3e176a9e641e758b1ba64c9560991/

And many more.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

python3: move toolchain comment above main option

So suboptions are correctly indented.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

python3: add hash file

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

python3: bump to version 3.4.3

All patches are simply refreshed, except
002-properly-detect-if-python-build.patch which is removed because an
identical fix has been merged upstream:

  https://github.com/python/cpython/commit/e128ea78e5964929ca902bebc07c242d58609dfc

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

vlc: uclibc-ng can be used

Tested with X86 and ARM compiles.

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/dvdrw-tools: needs host-m4

Fixes
http://autobuild.buildroot.net/results/b2b/b2b836b095251950e3229b66d6ca9faeaec45e72/
http://autobuild.buildroot.net/results/03b/03be24979d3ec4b08ebc7841ce1e2bcb05d26f97/
http://autobuild.buildroot.net/results/083/083ce6e998154a77dbaa16b3bed08570b745e587/
http://autobuild.buildroot.net/results/a51/a51da652314b01ae15f99708c5551f571247342b/

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

manual: don't use a comment at the end of the version variable assignment

The comment at the end of the version string assigment could lead to a download
failure so don't use it as a example.

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

package/vlc: libX11 is an optional dependency for the qt4 module

Fixes
http://autobuild.buildroot.net/results/d15/d158095d4caa8f58f1d205aeeb1ae707a1311ea5/
http://autobuild.buildroot.net/results/343/343e2981b60d9d09a4fcd8e2109540f0dc11b402/

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

memcached: bump to version 1.4.24

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

gcc/4.8.4: fix aarch64 vmlaq_lane_s32 typo

Fixes webp build:
http://autobuild.buildroot.net/results/656/6563be62557ab6c1bdd4152523774316dc09c9ce/
http://autobuild.buildroot.net/results/e31/e31e782d927215dde87b2f3174d70b8eb70085fd/
http://autobuild.buildroot.net/results/e2e/e2e627dd6fdfa58ee07fd3aaca43e1731cc8b6e4/
http://autobuild.buildroot.net/results/26e/26ed07fe917e1af21b5c5c3e4f6c2b39d78a8aff/

And many more.

Upstream bug report and patch:
https://code.google.com/p/webp/issues/detail?id=230
https://android-review.googlesource.com/#/c/99470

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ramspeed: add hash

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

memtester: add hash

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

texinfo: add hash file

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

spawn-fcgi: add hash file

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

sed: add hash file

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

gstreamer: add hash file

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

gst-plugins-ugly: add hash file

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

gst-plugins-good: add hash file

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

gst-plugins-base: add hash file

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

gst-plugins-bad: add hash file

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

gst-ffmpeg: add hash file

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

screen: add hash file

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

neon: add hash file

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

lzip: add hash file

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

libgpg-error: add hash file

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

heimdal: add hash file

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

guile: add hash file

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

diffutils: add hash file

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

civetweb: use TARGET_CONFIGURE_OPTS

Instead of explicitly handling CC/LD/CFLAGS. No functional change, but it is
cleaner like this.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

civetweb: rename lua option to match naming convention

Suboptions of package foo should be prefixed with BR2_PACKAGE_FOO, not just
BR2_FOO.

Also add Config.in.legacy handling for it.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

civetweb: unbreak static linking builds

Fixes
http://autobuild.buildroot.net/results/0a5/0a555a4ad0d100a7d402908b1f26d213de0ec127/
http://autobuild.buildroot.net/results/97b/97b014b2a611ba2528dda86bb05d6ffbec51c21d/
http://autobuild.buildroot.net/results/c2e/c2e314127be74a41bcdb78f04bc51fe2edef0b0f/
http://autobuild.buildroot.net/results/734/7346a89b5e7dd3633ecf9ac5292379ec9e832b2e/

And many more.

It is only the lua suboption that needs dlopen support, so adjust the
toolchain dependencies and CIVETWEB_LIBS handling to match.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

CHANGES: Update with recent changes

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/gnuchess: needs c++

Quoted from ChangeLog:

2013-11-23  Antonio Ceballos  <aceballos@gmail.com>
[...]
    * doc/gnuchess.texi: frontend no longer compiled in C; now in C++.

Fixes
http://autobuild.buildroot.net/results/04e/04e7c14c20f0e056e63a712be3821a3f5c94cff1/

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

freerdp: fix patch naming

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

package/nvidia-tegra23-codecs: new package

nvidia-tegra23-codecs provides various HW-accelerated media codecs, such
as: AAC3, h264, jpeg, MP3...

Signed-off-by: Nicolas Serafini <nicolas.serafini@sensefly.com>
[me:
  - split codecs from binaries into this patch
  - use our extractor helpers
]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[Thomas:
  - Remove <pkg>_DEPENDENCIES variable that contained
    nvidia-tegra23-binaries, since this dependency is a runtime
    dependency only.
  - Remove 'NVIDIA_TEGRA23_CODECS_INSTALL_STAGING = NO', since this is
    the default.
  - Remove 'NVIDIA_TEGRA23_CODECS_INSTALL_TARGET = YES', since this is
    the default.
  - Instead of using a NVIDIA_TEGRA23_CODECS_FIRMWARE variable listing
    the firmware files to copy, simply copy all firmware files.]

package.nvidia-tegra23-binaries: add gstreamer plugins

[Thomas:
  - add missing libXv dependency.
  - remove the NVIDIA_TEGRA23_BINARIES_GSTREAMER_PLUGINS and
    NVIDIA_TEGRA23_BINARIES_MULTIMEDIA_APPS, and use simple 'cp -dpfr'
    calls to copy the plugins and sample applications.]

Signed-off-by: Nicolas Serafini <nicolas.serafini@sensefly.com>
[yann.morin.1998@free.fr: split out into their own patch]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/nvidia-tegra23-binaries: new package

This patch adds a new package for all libraries and drivers
provided by Nvidia Linux4Tegra release 16.5.

We have intermediate .mk and Config.in, because those values will be
shared with the codecs package, to come in a follow-up patch, like we
have for the Freescale stuff.

Signed-off-by: Nicolas Serafini <nicolas.serafini@sensefly.com>
[yann.morin.1998@free.fr:
  - split jpeg into its own patch
  - split codecs and gstreamer plugins out into their own packages
  - do not patch for .pc files, just bundle them in $(@D)
]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ben Ben <carmazen84@gmail.com>
[Thomas:
  - don't select BR2_PACKAGE_XLIB_LIBXT, there is no reference to it
    in <pkg>_DEPENDENCIES, and it doesn't seem to be used by any
    library.
  - select BR2_PACKAGE_XLIB_LIBX11, which is referenced by the .mk
    file, and is actually used by the pre-built .so files.
  - do not select BR2_PACKAGE_XLIB_LIBXV, since it isn't used by the
    X.org libraries/drivers, but only the Gstreamer ones.
  - remove 'NVIDIA_TEGRA23_BINARIES_INSTALL_TARGET = YES', it is not
    needed since it is the default.
  - remove the definitions of NVIDIA_TEGRA23_BINARIES_FIRMWARE,
    NVIDIA_TEGRA23_BINARIES_LIBRARIES and
    NVIDIA_TEGRA23_BINARIES_PKGCONFIG, these are very long, and not
    really needed: just copy all the firmware, libraries and .pc files
    with a simple 'cp -dpfr'.]

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

package/xserver_xorg-server: add version 1.14.7

Some old binary blobs drivers, like the NVidia Tegra2/3 blobs, only
implement an old VIDEODRV ABI.

Since VIDEODRV ABIs are not backward compatible [0], lets introduce an
older Xserver version that supports such an old ABI.

[0] http://www.x.org/wiki/XorgModuleABIVersions/

[Thomas:
  - rename option BR2_PACKAGE_XSERVER_XORG_SERVER_V_1_14_7 to
    BR2_PACKAGE_XSERVER_XORG_SERVER_V_1_14
  - add hash file, change provided by Nicolas Serafini.]

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

package/xserver_xorg-server: not all versions need all dependencies

Support for the Present protocol has been added to, and is mandatory
since version 1.15.

As we're on the verge of adding 1.14.7, we need to make Present a
version-specific dependency.

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

package/xserver_xorg-server: add blind options to specify driver ABI

The upcoming NVidia Tegra2/3 binary blobs only work with a certain
VIDEODRV ABI (namely up to 14), so we must have a way to represent that
ABI so that such packages can not be enabled with a too recent/old ABI.

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

package/xserver_xorg-server: make the version a choice

Only a single entry for now, but we'll add one more in an upcoming
patch, to support old binary-only drivers, like the NVidia Tegra2/3
drivers.

[Thomas:
 - rename option from BR2_PACKAGE_XSERVER_XORG_SERVER_V_1_17_1 to
   BR2_PACKAGE_XSERVER_XORG_SERVER_V_1_17, since we don't want to
   rename the option at each minor release of the X.org server, and
   X.org servers with the same major version will not have a different
   ABI.]

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

package/xserver_xorg-server: prepare for multiple supported versions

Simply move patch around...

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

Makefile: add a few more targets to nobuild_targets

With this commit, one can now execute the source-check, external-deps
and legal-info targets regardless of the checks normally being done by
packages on the configuration.

Note that we intentionally do not go down the road of adding %-source,
%-legal-info, and the miryad of other targets that could work in such
situations. We only whitelist a few targets that are really useful to
have as nobuild_targets.

[Thomas: also add 'clean' and 'distclean' to the nobuild_targets, as
suggested by Yann.]

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

packages: refactor checks using BR_BUILDING

Instead of manually testing MAKECMDGOALS, use the newly introduced
BR_BUILDING variable to know if we're building or not.

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

Makefile: add BR_BUILDING variable

Some packages do some sanity checks on their configuration, for
example linux checks that the defconfig string is not empty when a
defconfig is used. Such checks are currently always performed, except
when the 'source' target is part of make goals.

This is problematic for two reasons:

 - Other targets such as 'source-check', 'external-deps' or
   'legal-info', that do not consist in doing a build, cannot be
   executed in such situations.

 - The current code removes the check as soon as one of the targets is
   source. But if there are other non-source targets called at the
   same time, the checks are ignored.

This commit therefore introduces an internal variable called
BR_BUILDING, which tells packages if we are actually building or
not. A variable nobuild_targets indicates the targets that we do not
consider as being build targets.

For the moment, nobuild_targets only contains 'source', to be
completely iso-functional.

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

Makefile: remove unneeded variables

Now that all the external-deps, source-check and source targets are
properly implemented based on the package infrastructure, the
PACKAGES_SOURCE, TARGET_HOST_DEPS, HOST_DEPS and HOST_SOURCE variables
are no longer needed. This is a good thing since they were anyway
incorrect, as they were only doing a two level recursion in the
dependencies of host packages.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[tested with a randpackageconfig]

Makefile: implement the 'source' target using the package infrastructure

Now that all the bits are in place, switch the global 'source' target
to use the package infrastructure logic.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[with 'make source' (actually together with the next patch).]
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[tested with a randpackageconfig]

pkg-generic: introduce a <pkg>_ALL_DOWNLOADS variable and factorize code

Now, both the download and source-check code are iterating over
<pkg>_SOURCE, <pkg>_PATCH and <pkg>_EXTRA_DOWNLOADS elements, figuring
out whether they contain full URLs or not. Instead of doing this
repeatdly, this patch introduces an internal <pkg>_ALL_DOWNLOADS
variable, which contains the list of everything that needs to be
downloaded, with URLs already expanded to take into account <pkg>_SITE
if needed.

This allows to simplify quite significantly the .stamp_download and
source-check implementation.

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

pkg-download: fix indentation for SOURCE_CHECK_* macros

Some of the SOURCE_CHECK_* macros are using a non-standard two-spaces
indentation. This commit switches them to use a single tab based
indentation, like in the rest of Buildroot.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[with git show -w]
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

pkg-download: get rid of DL_MODE

The DL_MODE variable is now no longer used with any other value than
"DOWNLOAD", so it no longer makes sense to have this variable at
all. Therefore, this commit gets rid of it.

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

pkg-generic: don't use DL_MODE in .stamp_downloaded

The .stamp_downloaded target is now only being used to really
download, and no longer for other activities like "source check" or
"external deps". So the check on DL_MODE being equal to DOWNLOAD is no
longer necessary.

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

pkg-generic: remove the .stamp_rsync_sourced fake stamp file

The only reason for the .stamp_rsync_sourced fake stamp file target to
exist was to handle the SOURCE_CHECK operation on packages using the
OVERRIDE_SRCDIR mechanism. Now that this is handled directly inside
$(1)-source-check, there is no longer any need for this part of the
code.

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

Makefile: implement a package based source-check target

This commit switches the implementation of the global source-check
target to use a package infrastructure based mechanism, using the
$(1)-all-source-check target added in the previous commit.

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

pkg-generic: implement source-check targets

This commits extends the common package infrastructure with the
$(1)-source-check and $(1)-all-source-check targets.

The $(1)-source-check target simply calls the newly added
SOURCE_CHECK macro on all items to be downloaded.

The $(1)-all-source-check target will depend on the
$(1)-all-source-check targets of all dependent packages and the
$(1)-source-check target of the current package, which allows to do a
recursive source-check in the dependency tree.

[Thomas: move the code around a bit to avoid repeating the
<pkg>_OVERRIDE_SRCDIR condition, as suggested by Yann.]

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

pkg-download: extend DOWNLOAD_INNER, add a SOURCE_CHECK macro

As part of moving to a package infrastructure based source-check
implementation, we are going to move away from the global DL_MODE
variable to select the behavior of the DOWNLOAD_INNER macro.

As a preparation to this, this commit makes the DOWNLOAD_INNER macro
take a third argument, which is the action to be done: either DOWNLOAD
or SOURCE_CHECK. For now, the DOWNLOAD macro passes $(DL_MODE) as this
third argument, in order to keep the existing behavior.

In addition, a SOURCE_CHECK macro is added, which calls DOWNLOAD_INNER
with the appropriate action. This macro will be used in the upcoming
package infra based implementation of source-check.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[tested by doing a "make source" on a randpackageconfig]
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

Makefile: move source-check outside of noconfig_targets

make source-check is here to check whether the remote sources for the
current selection of packages are still available. In its current
implementation, since it simply calls recursively a sub-make with the
source target, it can be a noconfig_targets. However, a follow-up
change will make source-check not use a sub-make, which will require
it to no longer be a noconfig_targets.

Therefore, as a preparation, this commit moves source-check outside of
noconfig_targets.

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

pkg-download: remove support for the SHOW_EXTERNAL_DEPS DL_MODE

Now that the external-deps implementation relies on the per-package
<pkg>-all-external-deps and <pkg>-external-deps targets and no longer
on the 'source' target with a custom DL_MODE, we can get rid of the
support for the SHOW_EXTERNAL_DEPS DL_MODE value in the pkg-download
logic.

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

Makefile: use the package infra based external-deps

This commit changes the global 'external-deps' target to use the newly
introduced per-package <pkg>-all-external-deps, instead of relying on
the 'source' target with a custom DL_MODE.

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

pkg-generic: take into account patch dependencies in source, external-deps and legal-info

The $(1)-all-{source,external-deps,legal-info} targets currently only
take care of the dependencies in <pkg>_DEPENDENCIES, but not
<pkg>_PATCH_DEPENDENCIES. This patch fixes that by introducing a
<pkg>_FINAL_ALL_DEPENDENCIES. It also reworks the $(1)-show-depends
target to make use of <pkg>_FINALL_ALL_DEPENDENCIES.

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

package/libsrtp: bump version to 1.5.2

Removed patches applied upstream:
https://github.com/cisco/libsrtp/commit/cae277e4377acc90ef787d09ed0d82968bbe809a
https://github.com/cisco/libsrtp/commit/b5fad5f893c7f323a8f8e1c512fa48cf6cdb1c8d

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

package/ccache: bump to version 3.2.1

Bump ccache from v.3.1.x to v.3.2.x.
All functionality remained the same.
 - Configure now accepts a parameter to use ccache's internal zlib.
 - conf.c was introduced and some settings were moved from ccache.c,
   with minor changes.
 - The PATCH_CONFIGURATION hook was moved from POST_CONFIGURE to
   POST_PATCH, as patching files should be done there, before calling
   the configure script.

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

libedit: add hash file

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

package/libedit: bump version to 20150325-3.1

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

package/tor: limit libseccomp support to archs supported by upstream

See this commit for a list of supported archs:
https://gitweb.torproject.org/tor.git/commit/src/common/sandbox.c?id=b0c1c700114aa8d4dfc180d85870c5bbe15fcacb

Fixes mips64el build error:
http://autobuild.buildroot.net/results/f97/f9792a0b013f6b79774ecc978d859f56bd941586/

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

uclibc-ng: remove temporary workaround

fminf and others got implemented in the time after this
patch was added. Remove the workaround to get mesa compiled
with uClibc-ng.

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/gnuchess: bump version to 6.2.1

Removed unneeded patches, Debian did the same:
http://metadata.ftp-master.debian.org/changelogs/main/g/gnuchess/unstable_changelog

gnuchess (6.0.2-1) unstable; urgency=low

  * New upstream version (Closes: #637722)
[...]
    - Drop patches

The patch is no longer needed because the gnuchess code has been
reworked. The input.c file no longer exists, and there is no longer a
getline() function being defined.

[Thomas: tweak commit log.]

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

xdriver_xf86-video-imx: new package

X.Org driver for iMX5 GPU. Most of patches come from Yocto project.

Tested using:
    mx5loco_defconfig +
    BR2_TOOLCHAIN_EXTERNAL=y
    BR2_ROOTFS_DEVICE_CREATION_DYNAMIC_EUDEV=y
    BR2_PACKAGE_QT5=y
    BR2_PACKAGE_QT5BASE_LICENSE_APPROVED=y
    BR2_PACKAGE_QT5BASE_EXAMPLES=y
    BR2_PACKAGE_QT5BASE_OPENGL_LIB=y
    BR2_PACKAGE_QT5BASE_XCB=y
    BR2_PACKAGE_QT5BASE_DEFAULT_QPA="xcb"
    BR2_PACKAGE_QT5BASE_GIF=y
    BR2_PACKAGE_QT5BASE_JPEG=y
    BR2_PACKAGE_QT5BASE_PNG=y
    BR2_PACKAGE_QT5QUICKCONTROLS=y
    BR2_PACKAGE_XORG7=y
    BR2_PACKAGE_XSERVER_XORG_SERVER=y
    BR2_PACKAGE_XSERVER_XORG_SERVER_MODULAR=y
    BR2_PACKAGE_XDRIVER_XF86_INPUT_EVDEV=y
    BR2_PACKAGE_XDRIVER_XF86_VIDEO_IMX=y
    BR2_PACKAGE_FREESCALE_IMX=y
    BR2_PACKAGE_FREESCALE_IMX_PLATFORM_IMX53=y
    BR2_PACKAGE_GPU_AMD_BIN_MX51=y
    BR2_PACKAGE_GPU_AMD_BIN_MX51_EXAMPLES=y

[Thomas: minor tweaks to Config.in.]

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/dovecot: Fix broken logic for comment display

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

fs: only add non rootfs- targets to PACKAGES

Commit 4570dbcb14b394d8f3c97baf4c9e5f4f225d1d37
("4570dbcb14b394d8f3c97baf4c9e5f4f225d1d37") added the dependencies of
rootfs building logic to PACKAGES, only thinking about the case where
rootfs building logic actually depends on real, normal packages
(host-mtd, host-genext2fs, etc.). But some file systems (rootfs-ubi)
depend on other filesystem images (rootfs-ubifs). And such targets
should not be added to PACKAGES, otherwise an incorrect circular
dependency is created.

This commit fixes that by only adding the rootfs building logic
dependencies that do *not* start with rootfs- to the global PACKAGES
variable.

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

qt: add QT_COORD_TYPE to double option

On ARM-based platforms, qreal is a typedef for float for performance
reasons. But we may want a double precision anyway.

Signed-off-by: Richard Genoud <richard.genoud@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-pyqt: fix PyQt_qreal_double enabling/disabling

fix :
http://autobuild.buildroot.net/results/6a2/6a28855c11b321ce8ceaf0acdd5395738af931fd/
http://autobuild.buildroot.net/results/91a/91a3641d1e4126475bbca0d3c779582832f6db91/
http://autobuild.buildroot.net/results/b10/b10fa70c199fc2de405068fea1eac80c29577747/
http://autobuild.buildroot.net/results/b53/b5334ac80afb58a19bd40b7c0b18378d75bc8fc7/
http://autobuild.buildroot.net/results/01f/01fa1319e467c3c2410fd462dc40c18b82dc8246/
http://autobuild.buildroot.net/results/2e6/2e6cae93ab860175c405e440bc7e24334b5b14e8/

According to configure-ng.py, PyQt_qreal_double must be disabled on
ARM target and when QT_NO_FPU is set.

[Thomas: refactor code using a hidden Config.in boolean.]

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

package/perl-encode-detect: new package

[Thomas:
 - rewrap Config.in help text
 - fix license, it's MPL-1.1, not 'open_source'.]

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

package/perl-db-file: new package

[Thomas: rename package to perl-db-file.]

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

package/tor: new package

[Thomas:
  - remove reference to BR2_LARGEFILE, this option has been
    removed. Largefile support is now guaranteed to be present.
  - rewrap Config.in help text to the appropriate length.]

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

package/perl-libwww-perl: Move package/perl-libwww to new package

This move is necessary to sync the package with scancpan output
as suggested by Francois.
http://lists.busybox.net/pipermail/buildroot/2015-April/126406.html

Since perl-libwww was never part of any official buildroot release
I did not bother the include legavy handling of its options ;)

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

package/input-tools: add hash for joystick_20051019-5.diff.gz

Fixes:
http://autobuild.buildroot.net/results/1f3/1f3c5c060cd98a3f4cf9688fc8a83cc1c23ece29

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libevdev: bump version to 1.4.2

Update version and hash (see [1] for details).

[1] http://lists.freedesktop.org/archives/input-tools/2015-April/001167.html

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/download: restore mandatory check of hashes

Now that custom external toolchains to be downloaded properly instruct
to not fail on a missing hash, restore the mandatory hash check for
everything else.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/uclibc: ignore missing hash for snapshots

We can't have a hash for snapshots, so add them to the exclusion list.

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>
Cc: Arnout Vandecappelle <arnout@mind.be>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain/external: ignore missing hash for custom downloaded toolchain

We will *always* be missing a hash file for custom external toolchains
that are downloaded.

So, just ignore that failure.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/download: add possibility to not fail on missing hash

In very constrained cases, it might be needed to not fail if a hash is
missing. This is notably the case for custom external toolchains to be
downloaded, because we do have a .hash file for external toolchains,
but we obviously can not have hashes for all existing custom toolchains
(he, "custom"!).

So, add a way to avoid failing in that case.

>From the Makefile, we export the list of files for which not to check
the hash. Then, from the check-hash script, if no check was done, and
the file we were trying to match in in this exclusion list, we just exit
without error.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
changes v6 -> v7:
  - /beautify/ the pattern in the case clause

Changed v5 -> v6:  (Arnout)
  - fix the pattern in the case clause

Changes v4 -> v5:
  - micro-optimisation, use case-esac instead of a for-loop  (Arnout)
  - typoes  (Arnout)

Changes v3 -> v4:
  - drop the magic value, use a list of excluded files  (Arnout)

Changes v1 -> v2:
  - fix typoes in commit log

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>