uclibc: fix tests building in uClibc 0.9.33.2

Following patches are required for successful bulding of uClibc built-int tests
(BR2_UCLIBC_INSTALL_TEST_SUITE=y):

1.
uclibc-0059-test-cater-for-config.patch
http://git.uclibc.org/uClibc/commit/test?id=7598eeaa1defa2884adaa890bb115c493d69cc35

2.
uclibc-0058-test-tls-fix-build-with-newer-binutils.patch
http://git.uclibc.org/uClibc/commit/test?id=931e8391565323ed2e589c83b83a7345813a5514

3.
uclibc-0060-test-Fix-math-c-dependency.patch
http://git.uclibc.org/uClibc/commit/test/math?id=a20a91ad7c042c46e4a2adee6d03315f857f9985

All mentioned patches must be removed as obsolete on bumping uClibc to the next
release.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Anton Kolesov <akolesov@synopsys.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Makefile: do not purge locale-archive from target

The rules to purge unwanted locales from the target also removed
the locale-archive file from /usr/lib/locale which contains the
locale definitions explicitly generated for the target.

Signed-off-by: Sven Neumann <neumann@teufel.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

makedevs: add license information

[Thomas: change license to be GPLv2. The header of makedevs.c is a bit
unclear, as it mentions both the GPL and LGPL. But since the code
seems to originally be coming from Busybox, assuming GPLv2 is the
safest option.]

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

infra: fix "<pkg>-rebuild" when using rsync source feature

Because the configure stamp file depends on the rsync stamp file,
removing it by the "<pkg>-clean-for-rebuild" rule trigger the configure
step. To avoid triggering the configure step use an order-only dependency.

Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Reported-by: Cédric Marie <cedric.marie@openmailbox.org>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

qt5webkit-examples: new package

[Thomas: remove QT5WEBKIT_EXAMPLES_INSTALL_STAGING = YES.]

Signed-off-by: Massimo Callegari <massimocallegari@yahoo.it>
Signed-off-by: Fatih Aşıcı <fatih.asici@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs: add new configuration for qemu-system-xtensa

Tested with Qemu 2.0.0.
Overlay is from
https://github.com/jcmvbkbc/xtensa-toolchain-build/tree/master/overlays

Many thanks to Max Filippov <jcmvbkbc@gmail.com> for his extended help
and kernel config examples.

A small kernel patch is required, because mkimage does not contain
support for Xtensa and the uImage target is failing. Redboot is unused
so disable it, too.

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

linux: add option to explicitly specify the kernel image name

For example the upcoming qemu-xtensa patch is using this feature,
where the target is called "zImage", but the resulting kernel name
is "Image.elf".

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

at: add license information

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

exim: allow using a custom configuration file

exim has lots of options configurable before the build process. It's not
useful to have all of them exposed in Buildroot, yet users may need to tweak
them.

Allow the user to pass an entire configuration file to make exim totally
reconfigurable without bloating the Buildroot menus.

[Thomas: remove "for exim" from Config.in options prompts.]

Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

lm-sensors: sensors-detect requires perl

The sensors-detect program is in fact a Perl script, so it won't run
on the target unless Perl is enabled in the configuration.

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

lm-sensors: don't use host ldconfig

To know whether the libiconv library is available, lm-sensors is using
the host ldconfig, which is obviously wrong in cross-compilation.

Moreover, the lm-sensors program making use of the iconv_*() API
already does it when __UCLIBC__ is *not* defined. In this case, the
iconv_*() functions are already part of the C library, so there is no
need to link against a separate library. Therefore, this patch simply
removes the libiconv handling.

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

lm-sensors: fix static-only builds

This commit adds a patch to lm-sensors to support a new variable
BUILD_SHARED_LIB that allows to conditionally enable or disable the
build (and usage) of the libsensors shared library.

It also refactors the make variables in a variable called
LM_SENSORS_MAKE_OPT.

Fixes:

  http://autobuild.buildroot.org/results/06c/06c197263e4939d6adc4877e152f602a69df751d/

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

php: add FPM support

[Thomas: adjust commit title, as suggested by Gustavo.]

Signed-off-by: Jordi Llonch <jordi.llonch@rochsystems.com>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

dependencies.sh: add other Debian package names for 32 bits libraries

[Thomas: implemented Arnout suggestion.]

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

uclibc: don't install tools to staging when static

Don't try to install ld* tools to staging when it's a static build,
there's no point and it'll be broken once we build a purely-static
uclibc.

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

statserial: don't link against legacy curses

Link against libncurses instead of the legacy libcurses.

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

dcron: new package

[Thomas: wrap help text, add BR2_USE_MMU dependency, minor fixes in
.mk file.]

Signed-off-by: Alvaro G. M <alvaro.gamez@hazent.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python, python3: build with tzset support

This patch adds support for time.tzset() in Python 2.7 and Python 3.

The Python configure script sets ac_cv_working_tzset to 'no' when
cross compiling, although it should be available and work with all
toolchains supported by Buildroot (?).
The patch below pre-sets the variable so that time.tzset() works.

[Thomas: adjust commit log.]

Signed-off-by: Henry Margies <henry.margies@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

bzip2: install shared library for the host

libbz2.so was explicitly built for the host, but was not installed. Install it.

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

python-pyzmq: enable PyZMQ for Python 3

The PyZMQ package is currently limited to Python 2; however, its
build system does not require the Python 2 interpreter and will work
as is with Python 3 (tested).

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

grub2: fix a few minor issues in help text

The help text of grub2 explains the detailed steps to create a disk
image with grub2 installed on it. However, the steps for the
BIOS-based systems have a few minor issues fixed by this patch:

 - When calling partx to get the partitions detected, we should do it
   on the /dev/loop0 block device, and not on the underlying disk.img
   image file.
 - The grub-bios-setup utility must be called as root to work properly
   on /dev/loop0.
 - The steps to cleanup the partx and loop device were missing.

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

binutils: remove 2.21.1

After the following commits:

26974bf6a7c94c0c44722fe072182cfcc49bdb6e use default binutils for sh architecture
6f1cf344d90ea3eabf292b04d993a7be2a785dd6 binutils: Use the default version by default on MIPS

the 2.21.1 binutils version is no longer required, and since it's an
old version, we get rid of it.

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

binutils: remove leftover version definitions

Following d4839ffcab70f2496673e5b376f4c0d492260396 ("binutils: remove
2.20.1, 2.21 and 2.23.1"), Config.in.host contains unneeded
definitions of BR2_BINUTILS_VERSION for no longer existing
versions. This patch fixes that.

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

perl-net-ssleay: add comment to new hook

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

uclibc: replace custom kconfig helpers with those provided by pkg-utils

This patch removes the custom kconfig helpers UCLIBC_OPT_SET and
UCLIBC_OPT_UNSET with the common KCONFIG_SET_OPT, KCONFIG_ENABLE_OPT and
KCONFIG_DISABLE_OPT.

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

pkg-utils: kconfig helpers: use single iso double quoting

The echo statements in the kconfig helpers are currently using double
quotes. For KCONFIG_SET_OPT this is problematic when the value argument
itself contains a double quote (a string value). In this case, the statement
    echo "$(1)=$(2)" >> $(3)
would become:
    echo "FOO="string value"" >> /some/path/.config
resulting in the string
    FOO=string value
in the config file, rather than the properly quoted
    FOO="string value"

The linux package worked around this by escaping the quote characters, but
a prettier solution is to use single quoting in the helpers (or
alternatively use no quoting at all).
A side effect of this change is that a $variable in the key or value would
no longer be interpreted by the shell, removing any unexpected behavior.

This change is only really necessary for KCONFIG_SET_OPT, but for symmetry
reasons the other helpers are updated too.

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

pkg-utils: kconfig helpers: add basic usage documentation

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

package/python: fix sqlite3 extension without threads

Fix a compiler error when building the sqlite3 extension, when the
toolchain does not have threads.

Fixes:
        http://autobuild.buildroot.net/results/10f/10f268cd55f16ee53f4c5e17e1248b0a1820f6ea/

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

perl-net-ssleay: fix configure

When TARGET_LDFLAGS contains more than one option, the generated
Makefile contains this definition of LDDLFLAGS:
        LDDLFLAGS = '--option-1 --option-2'

Which when passed to gcc is interpreted as a single option, and gcc
(rightfully) barfs on it, like with:
        arm-linux-gnueabihf-gcc: error: unrecognized command line
        option ‘-shared -g’

This is because perl-net-ssleay's buildsystem is really completely
brain-damaged. Other perl extensions do not behave like that, and
instead are doing the only sane thing to do: not add single quotes
around the definition, and just use what they were provided for
LDDLFLAGS.

So, just do the same for perl-net-ssleay. Since tweaking (yet once
more) their buildsystem is too complex, we just use a post-configure
hook to fix up the mess, by removing single quotes around the definition
of LDDLFLAGS.

Note: if only one option is specified, no single quotes are added,
but our hook is a no-op in this case.

See also:
    http://lists.busybox.net/pipermail/buildroot/2014-July/101782.html
and the previous message for even more entertainment. ;-)

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
[me: find the real cause of the issue, tweak the sed expression to not
 force -shared and instead just trim the single quotes, enhance commit
 log]
[Thomas: fix minor typos in the commit log, add comment above hook
 definition]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

polarssl: programs need MMU

Fixes:
http://autobuild.buildroot.net/results/b53/b535dfda85c8a25c5192c4be7540c4e852fce717/

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

nano: bump to version 2.3.5

Switch to the full-blown install when tiny is not selected, we really
want all of it in that case.

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

nasm: remove unused target package

When nasm was added in da8d09b716, it got both the host and the target
variants even though it doesn't have a Config.in and the target variant isn't
used anywhere. Therefore, this patch removes the unused target variant.

Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/perl-xml-libxml: not available for static link

Perl extensions are loaded at runtime with dlopen(), so it does not
make sense to even build extensions that are written in C.

perl-xml-libxml is written in C, so it is not available when doing a
static build.

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

package/perl-net-ssleay: not available for static link

Perl extensions are loaded at runtime with dlopen(), so it does not
make sense to even build extensions that are written in C.

perl-net-ssleay is (partly) written in C, so it is not available when
doing a static build.

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

package/perl-net-ssleay: fix buildsystem for cross-compilation

The buildsystem for perl-net-ssleay is totally brain-damaged: it tries
to execute the openssl program. It uses that to check the openssl
version, detect the prefix to openssl, and set the include and library
search paths.

Of course, it does not handle the destdir idiom, so it ends up finding
the host system's openssl (not even ours!), so it sets search paths to
/usr/inlcude and /usr/lib and /lib.

Well, guess what? It does not work.

Fix that by:
  - passing the prefix to openssl in the environment, so it does not go
    hunting for the host-system openssl;
  - not checking the version string, since we can't run, and Buildroot
    has the correct version anyway;
  - not setting any -I or -L paths since our compiler/linker already
    know where to look for for includes and libs.

Fixes:
    http://autobuild.buildroot.net/results/135/135867ef85535863e3647cc5fb82accb6f77612c/

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
[me: enhance commit log, add patch to avoid running openssl and setting
 wrong include and library search paths]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

infra/pkg-perl: add possibility to pass extra env at configure time

Some perl packages may use environment variables as a hint to know how
to be configured.

That's for example the case for perl-net-ssleay that uses
OPENSSL_PREFIX, if it is set in the environment, as the prefix to
openssl.

Add a new variable that packages can set if they need extra environment
variables. Update the manual accordingly.

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[yann.morin.1998@free.fr: enhance the commit log]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

busybox: fix typo 'overwritting'

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

ncmpc: bump to version 0.23

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

libgpgme: disable all tests

Disable all of the tests, including runtime that needs gnupg.
There's no longer --without configure options for them, it's now
--disable. Fixes:
http://autobuild.buildroot.net/results/eb1/eb1e06b1d99c441b6086c07ee04b084a61041dc7/

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>

binutils: fix xtensa ld segfault on linking linux modules

This fixes the following build error reproducible with any linux kernel
module built with a recent buildroot toolchain:

  LD [M]  fs/jbd2/jbd2.ko
  /bin/sh: line 1: 12069 Segmentation fault
  buildroot/host/usr/bin/xtensa-buildroot-linux-uclibc-ld -r --no-relax
  -T linux/scripts/module-common.lds --build-id
  -o fs/jbd2/jbd2.ko fs/jbd2/jbd2.o fs/jbd2/jbd2.mod.o

Backported from: e7d17e71cdc10a2e81e454ce3b9637f1b2a587f2
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gd: switch to new homepage

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

gettext: bump to version 0.19.1

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

linux-headers: bump 3.2.x series

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

busybox: fix interaction with version selection removal and lzo fix

The fix for the LZO issue in Busybox was merged between the moment the
patch removing the Busybox version selection was posted and
merged. This patch adjusts the Busybox patches to take into account
this issue: it removes the LZO patches for the no-longer existing
Busybox versions, and moves the LZO fix for the 1.22.1 Busybox version
as appropriate.

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

cifs-utils: bump to version 6.4

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

flashrom: bump to version 0.9.7

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

libgpgme: bump to version 1.5.0

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

libgcrypt: bump to version 1.6.1

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

polarssl: security bump to version 1.2.11

Fixes CVE-2014-4911 and a few other issues that don't have a CVE assigned
(backports from 1.3.x branch).
The no programs & shared/static patches are now upstream albeit in a
slightly different form.

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

e2fsprogs: bump to version 1.42.11

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

arch: avr32 should only get removed for the 2015.02 release

As discussed on the list.

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

qt5base: fix typo

Signed-off-by: Hadrien Boutteville <hadrien.boutteville@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

inadyn: not available for static

inadyn isn't static-build friendly, even if we pull in the libz link
line for the failure stated below it'll fail when trying to link to
libdl and using dlopen for the plugins. Fixes:
http://autobuild.buildroot.net/results/1a1/1a1a184e8248e5c51fa730d0c9721fe67a9c84a4/

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

atftp: fix static linking

Readline uses ncurses hence pull it in, also needs to explicitly link
against libpthread. Fixes:
http://autobuild.buildroot.net/results/259/2592612be38a2c1dde21b44be2ecf25d1cc5211a/

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

arch: deprecate the AVR32 architecture

The AVR32 architecture from Atmel is obsolete since a long time, not
supported upstream in most of the toolchain components (requires a
special version of gcc, an old version of uClibc, etc.). Until
February, Simon Dawson was making an excellent job at maintaining
AVR32 in Buildroot, and fixing all the issues caused by this
architecture. However, Simon focus has changed, and despite his call
for a new maintainer for AVR32 in Buildroot, nobody stepped up.

The issue of maintaining AVR32 is becoming worse and worse, so this
patch proposes to deprecate it for 2014.08, but keeping the support
around, to remove it for sure in 2014.11.

Cc: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

busybox: support only one version

This commit removes the version selection for the busybox
package. Busybox is very well maintained, and bugs are typically fixed
in a timely fashion. Moreover, regressions are fairly unlikely in this
very stable and well-tested tool.

Therefore, there isn't a very compelling reason to have a version
selection for Busybox since we don't accept such a version selection
for the vast majority of other packages, unless there is a strong
reason to do so.

Consequently, this commit:

 * Removes the 1.19.4, 1.20.2 and 1.21.1 Busybox versions, patches and
   default configuration file.

 * Moves the 1.22.1 patches from package/busybox/1.22.1 to just
   package/busybox/ like all other packages.

 * Renames the default 1.22.1 configuration file to just
   busybox.config.

 * Adapts the busybox.mk makefile to encode the current version to
   use.

 * Adds appropriate options to Config.in.legacy. However, even though
   the BR2_BUSYBOX_VERSION_1_22_X is removed, we don't add a
   Config.in.legacy option for it, since it would cause a legacy
   warning for virtually *all* users as most people are currently
   using 1.22.x.

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

gdb: remove versions 7.4 and 7.5

We already default to 7.6 for all architectures (except AVR32, ARC and
Microblaze that have their specific versions), and we have added 7.7
recently.

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

qt: remove gcc 4.6.x specific kludge

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

gcc: remove versions 4.3.x and 4.6.x

Those gcc series are old and are not used as the default versions for
any of the architectures we support, so this commit gets rid of them.

The gcc 4.3.x series technically remains used by the LPC32xx
defconfigs we have:

configs/ea3250_defconfig:BR2_GCC_VERSION_4_3_X=y
configs/fdi3250_defconfig:BR2_GCC_VERSION_4_3_X=y
configs/phy3250_defconfig:BR2_GCC_VERSION_4_3_X=y

Back when those defconfigs were introduced, gcc 4.3 was chosen because
it was the only one capable of building a fully working kernel for
those ARM-based platforms. However, the original submitter, Alexandre
Belloni, no longer has access to the hardware platforms, so he is
unable to test if newer gcc versions have fixed the problem. It
certainly doesn't make sense to keep gcc 4.3.x just for those three
boards, so we'll wait for someone actually using those defconfigs to
complain.

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

uclibc: remove BR2_UCLIBC_ARM_TYPE

The BR2_UCLIBC_ARM_TYPE was only used for uClibc 0.9.32 on ARM. Now
that only uClibc 0.9.33 is supported for ARM, there is no point in
keeping BR2_UCLIBC_ARM_TYPE, since the corresponding uClibc options no
longer exist.

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

package: remove references to uClibc 0.9.32

This commit removes all remaining references to uClibc 0.9.32.

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

uclibc: remove version 0.9.32.1

This commit removes the 0.9.32.1 version of uClibc, which is very old,
and does not bring any specific advantage over 0.9.33, which has been
around for more than two years now.

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

binutils: remove 2.20.1, 2.21 and 2.23.1

 - The 2.20.x series is old, it no longer makes sense to support it.

 - In the 2.21.x series, we still have 2.21.1, so keeping 2.21 doesn't
   make much sense, so this patch removes it.

 - Similarly for the 2.23.x series, having both 2.23.1 and 2.23.2
   doesn't make much sense, so this patch removes 2.23.1 and keeps
   2.23.2.

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

use default binutils for sh architecture

Tested toolchain compiles with binutils 2.22 and 2.24.
Result tested in Qemu 2.0.0. (only little endian version
tested)

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mpd: add twolame support

Now that there's a twolame package we can add support for it in mpd.

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

mpd: add pulseaudio deps comment, stylize libsndfile

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

mpd: remove redundant depends

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

toolchain-external: remove experimental Blackfin toolchain

The 2014R1 toolchain is provided in two flavours by Analog Devices:
one based on gcc 4.3.x, which is considered stable, and one based on
gcc 4.5.x, considered experimental. In commit
5a65b8e185fab38824da5e9b7b5b9e781e31d368 ("toolchain-external: add ADI
Blackfin 2014R1 toolchain, remove 2012R1") both variants were added.

However, after some testing in the autobuilders, and discussion with
the Analog Devices folks, it turns out that the experimental version
of the toolchain is too experimental. It causes numerous build
failures, and the Analog Devices folks clearly say that it's an early
release and that they expect quite a few problems to show up.

Therefore, this commit removes the experimental flavor and keeps only
the stable variant. Note that the removal/renaming of the Config.in
options is not a problem, since those options were added after the
2014.05 release.

Fixes:

  http://autobuild.buildroot.org/results/2a9/2a9d9c332a206fdb46bc8ba022c74d23082a6312/
  http://autobuild.buildroot.org/results/e1c/e1ce0c1cdd0139208dddaa8f2441ab0e3ab2385e/
  http://autobuild.buildroot.org/results/993/993aca3f4719afaa4b37524f9136fb8cdc53a066/
  and more.

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

linux: bump default to version 3.15.5

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

linux-headers: bump 3.{4, 10, 14, 15}.x series

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

rsync: bump to version 3.1.1

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

ccrypt: new package

ccrypt is a utility for encrypting and decrypting files and streams.

Signed-off-by: David Bachelart <david.bachelart@bbright.com>
Reviewed-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libv4l: Bump version to 1.0.1

Quite some work has been done from 0.8.3 to 1.0.1, and as a result
this commit is very intrusive. The biggest change is the move to an
autotools package.

Then, the options that enable utilities individually have been deprecated
and moved to Config.in.legacy. Instead, we introduce new option to select
either all the utilities. This change loses granularity in favor of
maintainability.

[Peter: needs host-pkgconf. Select and use argp-standalone on uClibc]
Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xserver_xorg-server: drop host-xfont_font-util dependency

This is no longer needed now that e244eb12d (xserver_xorg-server: Fix race
condition when installing man pages) has been reverted.

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

ntpd: drop default /etc/default/ntpd file

With the recent change to the init script the default /etc/default/ntpd file
doesn't do anything, so don't install it.

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

Fixing ntpd time syncing problem

Add ntp.conf file to make ntpd syncing.
Starting ntpd daemon with -g to sync time also with big offsets.
Removes the use of deprecated ntpdate command for initial time sync.

[Peter: drop unused NTPDATE_BIN variable]
Signed-off-by: Angelo Compagnucci <angelo.compagnucci@gmail.com>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openvpn: disable plugins for static builds

Fixes:
http://autobuild.buildroot.net/results/082/08295cdcb3d14198bc7fbddec89d6fead946afba/

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

cryptodev-linux: fix license information

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

cloop: drop unused target variant

We only support cloop for the host (for fs/cloop), so drop the unused target
variant.

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

cloop: add license information

[Peter: clarify that advancecomp is GPLv2+, add COPYING]
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

bootutils: add license information

[Peter: also add _LICENSE_FILES]
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libnl: Move the pthread dependency menu comment

In order for this comment to be properly displayed, it must be moved to be
at the end of the file.

Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xserver_xorg-server: Fix race condition when installing man pages, 2nd try

Should fix the still present bug:
http://autobuild.buildroot.net/results/045/04514bc9a8d1aad8ddd7af4f0fd8c16edd2e89a9/
http://autobuild.buildroot.net/results/1f2/1f25f744771fbb5124dca4563d79ea92cbfd72a3/

This patch reverts my previous attempt to fix this hard-to-catch race condition
http://git.buildroot.net/buildroot/commit/package/x11r7/xserver_xorg-server?id=e244eb12dfc50bb56167d5fdaef3aac52b0ff6a3

and removes make target install-data from _INSTALL_STAGING_OPT, which was
introduced in 2008:
http://git.buildroot.net/buildroot/commit/package/x11r7/xserver_xorg-server/xserver_xorg-server.mk?id=d4e56ff4dc8a0528550eb8503956bdcddb2fe516

Comparing the content of STAGING_DIR with and without "make install-data" showed
no difference so I assume that is it no longer necessary to explicitly call this
make target as the normal install target takes care of everything.

The race condition occurs because "make install" and "make install-data" both
install usr/share/man/man1/Xserver.1

[Peter: don't remove host-xfont_font-util]
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mpd: fix build failures with gcc 4.9.x

As reported by Sagaert Johan in the mailing list mpd 0.18.x fails to
build with gcc 4.9.x
Patch status: sent upstream.

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

python3: improve isolation of build environment and error handling

This commit mainly fixes bug #7268 by improving the isolation of
the build environment by making sure some host environment header
paths do not leak into the target build. The investigation was done by
David <buildroot-2014@inbox.com>. This is done by not calling the
add_multiarch_paths() function of setup.py when we're cross-compiling,
a change made in the newly introduced
python3-012-dont-add-multiarch-path.patch.

In addition to this, another patch is added to make sure the build is
aborted when one of the Python module fails to build. This is done in
python3-013-abort-on-failed-modules.patch. Without this, the Python
setup.py script simply logs which module failed to build, but doesn't
abort, so it's hard to notice when there is a problem.

[Peter: slightly reword commit message]
Cc: David <buildroot-2014@inbox.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

python: improve isolation of build environment and error handling

This commit mainly fixes bug #7268 by improving the isolation of
the build environment by making sure some host environment header
paths do not leak into the target build. The investigation was done by
David <buildroot-2014@inbox.com>. This is done by not calling the
add_multiarch_paths() function of setup.py when we're cross-compiling,
a change made in the newly introduced
python-013-dont-add-multiarch-path.patch.

In addition to this, another patch is added to make sure the build is
aborted when one of the Python module fails to build. This is done in
python-014-abort-on-failed-modules.patch. Without this, the Python
setup.py script simply logs which module failed to build, but doesn't
abort, so it's hard to notice when there is a problem.

[Peter: slightly reword commit message]
Cc: David <buildroot-2014@inbox.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dmraid: add license information

[Peter; license is GPLv2+, drop LGPL, add LICENSE file]
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

lesstif: add license information

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

libraw: add license information

[Peter: also add README to _LICENSE_FILES]
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

linux-fusion: add license information

[Peter: license is GPLv2+]
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

musepack: add license information

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

opencv: add license information

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

quota: add license information

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

setserial: add license information

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

texinfo: add license information

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

torsmo: add license information

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

usbmount: add license information

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

dvb-apps: fix static only build

This commit makes dvb-apps behave properly in a static-only build. It
passes the static=1 variable to the Makefile, and adds a patch that
fixes the Makefile to not build .so libraries when static=1.

Fixes:

  http://autobuild.buildroot.org/results/3db/3dbfa7b692d4b123b3ba6b55c84df12818273554/

[Peter: drop V=1 as pointed out by Yann]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

support/download: fix the git helper

When switching the git helper over to a shell script, a special case was
not carried over: in case the remote has the required reference, we
attempt a shallow clone, using --depth 1. However, this is not supported
when the remote is accessed with the http protocol.

Therefore, the download fails.

What happened before the conversion to a shell script was that the helper
in the Makefile would fallback to doing a full-clone.

This is the case and behaviour that were lost in the conversion.

To avoid making the script too complex, we only attempt a full clone if
needed. And we decide that a full clone is needed by default; we decide
it is unnecessary if the remote has the needed reference *and* the
shallow clone was successful.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

legal-info: depend on patch rather than extract

Switch legal-info generation dependency to patch rather than extract
so we can use debian/copyright from debian-maintained packages,
because the *orig* file alone doesn't contain the appropiate file.

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