support/testing: drop explicit docker kernel options from docker test

Since commit 0390777bfaff6c6 (package/docker-engine: needs some kernel
options), docker-engine now automatically ensures the needed kernel options
are enabled, so drop the explicit options from the kernel config.

23:19:27 TestDockerCompose                        Starting
23:19:28 TestDockerCompose                        Building
00:14:41 TestDockerCompose                        Building done
00:15:30 TestDockerCompose                        Cleaning up
.
----------------------------------------------------------------------
Ran 1 test in 3362.784s

OK

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

support/testing: drop explicit CGROUPFS_MOUNT from docker test

Since commit 4f8229653 (package/docker-engine: needs more runtime
dependencies), docker-engine now automatically pulls in cgroupfs-mount, so
drop the explicit handling of it in TestDockerCompose.

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

package/feh: bump version to 3.4.1

Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/webkitgtk: security bump to version 2.28.4

This is a minor release which provides fixes for CVE-2020-9862,
CVE-2020-9893, CVE-2020-9894, CVE-2020-9895, CVE-2020-9915, and
CVE-2020-9925.

Full release notes can be found at:

  https://webkitgtk.org/2020/07/28/webkitgtk2.28.4-released.html

A detailed security advisory can be found at:

  https://webkitgtk.org/security/WSA-2020-0007.html

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/wpewebkit: security bump to version 2.28.4

This is a minor release which provides fixes for CVE-2020-9862,
CVE-2020-9893, CVE-2020-9894, CVE-2020-9895, CVE-2020-9915, and
CVE-2020-9925.

Full release notes can be found at:

  https://wpewebkit.org/release/wpewebkit-2.28.4.html

A detailed security advisory can be found at:

  https://wpewebkit.org/security/WSA-2020-0007.html

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/kodi: bump version to 18.8-Leia

Removed patch which was applied upstream:
https://github.com/xbmc/xbmc/commit/920cc74da069216d19f724ecf5246cf997861bc6

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

package/dbus: disable systemd for host build

This fixes an issue if host-dbus happens to be rebuilt after systemd, in
which case it autodetects systemd support but then ignored the prefix
when installing unit files.  That means that is tries to write to the
host system's /usr/lib/ which fails.

There is no reason to build and install systemd support in the host
build, so disable it explicitly.

Signed-off-by: John Keeping <john@metanate.com>
Tested-by: Adam Duskett <aduskett@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/easydbus: bump to version 0.1.0

Update indentation in hash file (two spaces)

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Reviewed-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/iprutils: bump to version 2.4.18.1

Update indentation in hash file (two spaces)

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

{linux, linux-headers}: bump 4.{4, 9, 14, 19}.x / 5.{4, 7}.x series

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

package/linux-headers: drop support for 5.6.x headers

The 5.6.x series is now EOL so remove the option and add legacy
handling for it.

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

package/linux-headers: drop support for 5.5.x headers

The 5.5.x series is now EOL so remove the option and add legacy
handling for it.

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

package/libcurl: fix build against gnutls with proxy disabled

Add upstream patch (#4) fixing build with gnutls when
BR2_PACKAGE_LIBCURL_PROXY_SUPPORT is disabled.

Patch #4 depends on #3 to apply so add this one as well.

Fixes:
http://autobuild.buildroot.net/results/31d7204869ff71319ea055688c919a646bfb200b/
http://autobuild.buildroot.net/results/f8d2fb919475cdff4a36ad93071048ee09193b98/
http://autobuild.buildroot.net/results/2f07a0ac1240a6040a3509d2ebf06906a31fd172/

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/lz4: simplify build/install logic

No functional change.

Rather than tracking both build targets and sub directory to run 'make
install' from, use a make loop for the staging/target install logic, similar
to how we are doing it to for the build step and directly use the Makefile
in the sub directory rather than the helper lib/lz4 targets in the top level
Makefile.

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

package/luaposix: select bit32 based on lua ABI version 5.1

Since the bit32 module is needed by all lua ABI 5.1 providers we can
select based on the lua ABI version instead of both lua 5.1 and luajit.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/solarus: depend on lua ABI version 5.1

Since solarus is compatible with lua interpreters that provide the
version 5.1 ABI we should depend on that instead of unconditionally
selecting luajit.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/collectd: depend on luainterpreter for lua support

Since collectd lua support is compatible with any luainterpreter
implementation we should depend on BR2_PACKAGE_HAS_LUAINTERPRETER
instead of unconditionally selecting BR2_PACKAGE_LUA.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
[yann.morin.1998@free.fr: use luainterpreter as dependency]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/sysdig: depend on lua ABI version 5.1

Since sysdig is compatible with lua interpreters that provide the
version 5.1 ABI we should depend on that instead of unconditionally
selecting luajit.

Use luainterpreter as the dependency, now that either lua or luajit
can be used.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
[yann.morin.1998@free.fr:
  - reorder dependencies in Config.in: package deps go after toolchain
    deps
  - keep the toolchain deps together in the comment
  - switch to using luainterpreter as a build dependency
  - split the now-long _DEPENDENCIES list
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/luainterpreter: gather the ABI version string defaults

Now that we have booleans to represent the Lua ABI version, and that
every Lua providers do select those, there is no longer any reason
for them to also handle the ABI version string anymore.

Move the defaults into the common luainterpreter.

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Cc: James Hilliard <james.hilliard1@gmail.com>
Cc: Francois Perrad <francois.perrad@gadz.org>

package/luainterpreter: add abi version options

Since a number of packages depend on specific ABI version that are
provided by lua and luajit packages we should have a kconfig variable
that those packages can depend on.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/prosody: don't depend on !lua 5.3

The current version of prosody is compatible with Lua 5.3.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/minetest: depend on LuaJIT

Luajit is a provider for luainterpreter. We can't select providers of
virtual packages; we can only depend on them.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
[yann.morin.1998@free.fr:
  - a lua interpreter is needed, so don't drop the dependency
  - make it a "depends on" rather than a "select"
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/efl: depend on LuaJIT

Luajit is a provider for luainterpreter. We can't select providers of
virtual packages; we can only depend on them.

Note also that it is not very clear whether the host and target variants
of EFL need to be built with the same lua interpreter. Today, this is
guaranteed as we inly use luajit in both cases. But there were issues
with lua 5.1 in the past, so stick to only using luajit.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
[yann.morin.1998@free.fr:
  - depend on luajit, not "any" luainterpreter
  - which keeps the host and target variants built with the same
    interpreter
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/systemd: bump version to 245.7

Also drop upstream patch

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/stm32mp157c_dk2: Bump Kernel and U-Boot

Bump kernel to version 5.7.10 and U-Boot to 2020.07

Signed-off-by: Pierre-Jean Texier <pjtexier@koncepto.io>
Tested-by: Bartosz Bilas <b.bilas@grinn-global.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/stm32mp157a_dk1: Bump Kernel and U-Boot

Bump kernel to version 5.7.10 and U-Boot to 2020.07

Signed-off-by: Pierre-Jean Texier <pjtexier@koncepto.io>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/freescale-imx/firmware-imx: refactor DDR firmware preparation

The logic doing the padding and concatenation of DDR firmware was
duplicated between the DDR4 and LPDDR4 cases. This commit refactors
this logic into a single function FIRMWARE_IMX_PREPARE_DDR_FW.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Tested-by: Stephane Viau <stephane.viau@oss.nxp.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/janus-gateway: bump version to 0.10.3

Also remove upstream patch.

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

Makefile: properly account for custom tags in BR2_VERSION_FULL

BR2_VERSION_FULL is currently defined as follows:

  BR2_VERSION_FULL := $(BR2_VERSION)$(shell $(TOPDIR)/support/scripts/setlocalversion)

This BR2_VERSION_FULL value then gets used as the "VERSION" variable
in the /etc/os-release file.

The logic of "setlocalversion" is that if it is exactly on a tag, it
returns nothing.

If it is on a tag + a number of commits, then it returns only
-XYZ-gABC where XYZ is the number of commits since the last tag, and
ABC the git commit hash (these are extracted from git describe).

This output then gets concatenated to BR2_VERSION which gives
something like 2020.05 or 2020.05-00123-g5bc6a.

The issue is that when you're on a tag specific to your project, which
is not a Buildroot YYYY.MM tag, then the output of setlocalversion is
empty, and all you get as VERSION in os-release is $(BR2_VERSION)
which is not really nice. Worse, if you have another non-official
Buildroot tag between the last official Buildroot tag/version and
where you are, you will get $(BR2_VERSION)-XYZ-gABC, but XYZ will not
correspond to the number of commits since BR2_VERSION, but since the
last tag that "git describe" as found, which is clearly incorrect.

Here is an example: you're on master, "make print-version" (which
displays BR2_VERSION_FULL) will show:

$ make print-version
2020.08-git-00758-gc351877a6e

So far so good. Now, you create a tag say 5 commits "before" master,
and show BR2_VERSION_FULL again:

$ git tag -a -m "dummy tag" dummy-tag HEAD~5
$ make print-version
2020.08-git-00005-gc351877a6e

This makes you believe you are 5 commits above 2020.08, which is
absolutely wrong.

So this commit simplifies the logic of setlocalversion to simply
return what "git describe" provides, and not prepend $(BR2_VERSION) in
the main Makefile. Since official Buildroot tags match official
Buildroot version names, you get the same output when you're on an
official Buildroot tag, or some commits above a Buildroot tag. An in
other cases, you get a sensible output. The logic is also adjusted for
the Mercurial case.

In the above situation, with this commit applied, we get:

$ make print-version
dummy-tag-6-g6258cdddeb

(6 commits instead of 5 as we have this very commit applied, but at
least it's 6 commits on top of the dummy-tag)

Finally, if you're not using a version control system, setlocalversion
was already returning nothing, so in this case, the Makefile simply
sets BR2_VERSION_FULL to BR2_VERSION to preserve this behavior.

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

gitlab-ci: check all defconfigs on every push

The defconfig check has been introduced by the previous
patch before the building each defconfig but those builds
are done every week or more.

Checking if a defconfig is valid can be done on every
push in the repository since it take few seconds.

This would allow to detect as soon as possible a problem
in a defconfig and eventually avoid breaking the build
while build testing all defconfig.

Introduce a new job template ".defconfig_check" in
gitlab-ci.yml.in and modify the generate-gitlab-ci-yml
to create a job for each defconfig to run the test.

Although, we could have used only one job to do all
tests, using one job per defconfig allow to identify
easily in gitlab which defconfig is falling.

Tested:
https://gitlab.com/kubu93/buildroot/pipelines/138331069
https://gitlab.com/kubu93/buildroot/pipelines/171223758

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

gitlab-ci: check generated config files

Use the script added by the previous patch to check
generated config files.

Tested on gitlab:
https://gitlab.com/kubu93/buildroot/pipelines/137597966

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/scripts: add check-dotconfig.py

For the same reason as for 50b747f212be2c9c0f7cf10c674ed488d042715c,
we need to check if the generated configuration file (.config)
contains all symbols present in the defconfig file.

If not there is an issue with the defconfig.

This script will be used in .gitlab-ci.yml.

Inspired by is_toolchain_usable() function from genrandconfig:
https://git.busybox.net/buildroot/tree/utils/genrandconfig?h=2020.02#n164

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
[yann.morin.1998@free.fr:
  - strip defconfig lines when reading them
  - use a generator to read the defconfig lines
  - no need to strip() again when building the missing list
  - testing the list directly, not its len()
  - simply sys.exit(1) in the error condition
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

configs/rock_pi_n10_defconfig: remove BR2_TARGET_UBOOT_NEEDS_PYTHON

BR2_TARGET_UBOOT_NEEDS_PYTHON3 is already selected for
this defconfig, BR2_TARGET_UBOOT_NEEDS_PYTHON and
BR2_TARGET_UBOOT_NEEDS_PYTHON3 can't be selected at the
same time.

Reported by check-dotconfig.py.

Fixes:
https://gitlab.com/kubu93/buildroot/-/jobs/657885314

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

configs/rock_pi_4_defconfig: remove BR2_TARGET_UBOOT_NEEDS_PYTHON

BR2_TARGET_UBOOT_NEEDS_PYTHON3 is already selected for
this defconfig, BR2_TARGET_UBOOT_NEEDS_PYTHON and
BR2_TARGET_UBOOT_NEEDS_PYTHON3 can't be selected at the
same time.

Reported by check-dotconfig.py.

Fixes:
https://gitlab.com/kubu93/buildroot/-/jobs/657885313

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

configs/imx6ullevk_defconfig: remove typo

Reported by check-dotconfig.py.

Fixes:
https://gitlab.com/kubu93/buildroot/-/jobs/657885169

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/config-fragments: use external riscv64 musl toolchain

Move to an external RISC-V 64 bit musl toolchain to ease the load
on the autobuilders.

Signed-off-by: Mark Corbin <mark@dibsco.co.uk>
Reviewed-by: Romain Naour <romain.naour@gmail.com>
Tested-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

support/config-fragments: add RISC-V 64-bit uclibc autobuild config

Add a RISC-V 64-bit autobuild configuration for the internal
toolchain with uclibc.

Signed-off-by: Mark Corbin <mark@dibsco.co.uk>
Reviewed-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

toolchain/toolchain-buildroot: enable uclibc for riscv64

We can enable uclibc for RISC-V 64 bit now that it has been
bumped from v1.0.32 to v1.0.34.

Uclibc has had basic support for RISC-V 64 bit since v1.0.31, but
shared library and TLS/NPTL support has only been available since
v1.0.33.

This update has been tested using qemu_riscv64_virt_defconfig and
the Buildroot host QEMU.

Signed-off-by: Mark Corbin <mark@dibsco.co.uk>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/systemd: needs kernel options

according to https://wiki.gentoo.org/wiki/Systemd#Kernel

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
[Peter: drop optional CGROUP_BPF / BPF_SYSCALL, unused POSIX_MQUEUE and
 DEVTMPFS which is already taken care of by linux.mk]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/docker-engine: needs some kernel options

according to https://wiki.gentoo.org/wiki/Docker#Kernel /
docker-engine/contrib/check-config.sh.

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
[Peter: synchronize with contrib/check-config.sh, only enforce DM options if
 device mapper driver is enabled]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/lz4: don't install programs by default

Commit 2acb27e298c54935f890111fc5897a11350b913b ("package/lz4: make
program installation optional") make the installation of the lz4
programs optional. To preserve backward compatibility, this commit
added a "default y" to the BR2_PACKAGE_LZ4_PROGS.

While Buildroot definitely tries to preserve backward compatibility in
general, in this case, the backward compatibility would mean that
everbody would by default install those 220 KB of program, which are
most likely not needed for the majority of use cases. Installaing them
by default kind of breaks the Buildroot principle of minimalism.

In this specific case, it is believed that the principle of minimalism
wins over the principle of backward compatibility.

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

package/cegui: fix indentation

No functional changes, only whitespace changes.

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

package/cegui: fix selecting the default OpenGL renderer

According to the cegui requirements select explicitly the default
OpenGL renderer based on available dependencies.

Signed-off-by: Bartosz Bilas <b.bilas@grinn-global.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/lz4: make program installation optional

Often lz4 is used as a library, and not as a standalone program.
Excluding lz4 binary will save some space in this case. Indeed, the
lz4 program is always statically linked against its library, which
makes it duplicate the whole library size:

$ ls -l usr/lib/liblz4.so.1.9.2 usr/bin/lz4
-rwxr-xr-x 1 thomas thomas 226724 27 juil. 16:33 usr/bin/lz4
-rwxr-xr-x 1 thomas thomas 156996 27 juil. 16:33 usr/lib/liblz4.so.1.9.2

Since lz4 is now primarily a library, it's moved to the "Libraries"
section.

Of course, installation of programs defaults to "yes" to preserve
backward compatibility.

Signed-off-by: Ed Spiridonov <edo.rus@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-matplotlib: add QT5 support

python-matplotlib may be used in qt5 application. This patch add option to
enable qt5. Since this backend is automatically enabled when pyqt5 is detected,
list dependencies is simply an updated (no enable/disable options).

Signed-off-by: Gwenhael Goavec-Merou <gwenhael.goavec-merou@trabucayre.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/python-networkx: install for host-python3

This Python module only works for Python 3: the target variant has a
"depends on BR2_PACKAGE_PYTHON3".

However today in Buildroot, when no target python is selected, or when
BR2_PACKAGE_PYTHON=y, all host python modules are installed for
host-python, i.e Python 2. But this module won't install in Python 2,
so let's force its host variant to be installed with Python 3 on the
host. Of course, for that to work, its dependency must also be built
for host-python3, so we change it to the newly introduced
host-python3-decorator package.

Fixes:

  http://autobuild.buildroot.net/results/46a89bd6cd0d0b896b28010db287068309e7a43c/

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

package/python3-decorator: new package

The host-python-networkx package needs to be built for the
host-python3, even when the target Python is not necessarily Python
3.x. Since it depends on host-python-decorator, we need a Python 3
variant of it, which this patch introduces.

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

package/rocksdb: bump to version 6.10.1

- Update indentation of hash file (two spaces)
- Remove first patch (already in version)
- Add link to upstream PR in second patch
- Add upstream patch to fix build on powerpc (patch does not apply
  cleanly on current version)
- Update hash of README.md, slack forum added with
  https://github.com/facebook/rocksdb/commit/cf342464caf8155ab60f7a48e39a8d6ae3138a7a

Fixes:
 - http://autobuild.buildroot.org/results/242db7139ce5dd6ece4a1eb50a1aa47c841a41b7

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-rpi-ws281x: bump to version 4.2.4

Signed-off-by: Grzegorz Blach <grzegorz@blach.pl>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/{glibc, localedef}: bump to version 2.31

For glibc 2.31.x:

  - Update LICENSES file hash due to url change:
    "Prefer https to http for gnu.org and fsf.org URLs"

  - riscv64 does not build with kernel headers < 5.0, but upstream
    has not yet comitted a single fix, neither in master nor in the
    maintenance branch:
       https://sourceware.org/ml/libc-alpha/2020-02/msg00018.html

For localedef 2.31.x:

  - Remove upstream patch for localedef:
    0003-localedef-Use-initializer-for-flexible-array-member-.patch

Note that this version bump required some patches applied on
several packages (already applied):
    [Busybox] 13f2d688a24f47446af236829bd6ca194d5aea5b
    [openssh] bad75bca315dbd2c69f8a9cb02fa9f27636e3d48
    [gcc] disable libsanitizer with gcc 7.5

See:
https://sourceware.org/legacy-ml/libc-announce/2020/msg00001.html

Tested by toolchain builder:
https://gitlab.com/kubu93/toolchains-builder/pipelines/129551000

Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/gcc: disable libsanitizer with gcc 7.5 and glibc 2.31 toolchain

libsanitizer in gcc 7.x fails to build with glibc 2.31. Fixing it would
require backporting an upstream commit:

    https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=4abc46b51af5751d657764d0c44b8a4aeed06302

However, the backport is not trivial, as there are a lot of conflicts.

Disable libsanitizer since the gcc 7.5 branch is now closed
(unmaintained) and it's not a trivial merge.

Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/c-periphery: bump to v2.2.2

c-periphery now supports building on older kernels before 4.8 as there
are now checks for the new kernel cdev gpio interface.

Also updated hash file to two space format.

Signed-off-by: Ryan Barnett <ryanbarnett3@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/wine: bump version to 5.12

Also update hash file formatting (2 spaces).

Signed-off-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/setools: add comment about Python 3.x dependency

Commit 09b94b1e8f94d1f49f7a2349b4c8cbfecefa73da forgot to add the
usual comment that python3 dependency is due to python-networkx.

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/nodejs: use system-icu for host-nodejs

The nodejs configure.py file orders zlib headers before the bundled ICU
headers. The zlib headers happen to be located in the system include
directory, next to some system ICU headers (not bundled). If these are
built before nodejs is, nodejs will get confused and try to use the
system ICU headers instead of the bundled ones.

Fix this by always using host-icu.

Set CXXFLAGS to -DU_DISABLE_RENAMING=1 when building with
system-icu since host-icu is built with --disable-renaming.

Fixes:
 - http://autobuild.buildroot.net/results/1ef947553ec762dba6a6202b1cfc84ceed75dbb2/

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
[yann.morin.1998@free.fr:
  - keep alphabetical order in _DEPENDENCIES
  - don't introduce HOST_NODEJS_CONF_OPTS
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

gitlab: generate the gitlab-ci configuration before each build

Since Gitlab 12.9, Gitlab allow to trigger child pipeline with generated configuration file.
See: https://gitlab.com/gitlab-org/gitlab/-/issues/35632

This allow us to stop updating the .gitlab-ci.yml file when a
new defconfig is added to Buildroot.

Remove check-gitlab-ci.yml job since it is now uneeded.
Remove .gitlab-ci.yml make target.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
[ann.morin.1998@free.fr: manual: no longer needed to update at all]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/rust: drop reference to HOST_RUST_JEMALLOC_CONF

This variable is no longer used since
4e3be3ae9d4c9f1b0c035eee0c923a724f9a9316 ("package/rust: bump version
to 1.33.0")

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

package/cargo-bin: remove and use cargo-bin from rust-bin package

Rust upstream distributes tarballs with separate rust and cargo
binaries, but also tarballs with everything together. The latter
contains a bit more than what we need, so the download is slightly
larger than separate cargo-bin and rust-bin. But it simplifies our life
if we do the same in rust-bin as in rust (i.e. get both rust and cargo).

Signed-off-by: Patrick Havelange <patrick.havelange@essensium.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/rust: bump to 1.40

- remove of the llvm workaround patch as it is integrated in this
  newer version.
- bump cargo-bin to 0.41 (corresponds to rust 1.40) and update
  licenses hashes.
- bump rust-bin to 1.40.

Signed-off-by: Patrick Havelange <patrick.havelange@essensium.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: David Pierret <david.pierret@smile.fr>
Tested-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

docs/manual: cargo packages depend on host-rustc

As Cargo is either installed by host-rust or host-rust-bin, the
Cargo-based packages should depend on host-rustc.

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Patrick Havelange <patrick.havelange@essensium.com>
[Arnout: also select HOST_RUSTC in Config.in]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: David Pierret <david.pierret@smile.fr>
Tested-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/cargo: drop package

As no package depends on the standalone cargo package, it can be
dropped.

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Patrick Havelange <patrick.havelange@essensium.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: David Pierret <david.pierret@smile.fr>
Tested-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/ripgrep: depends on host-rustc

As all providers of rustc also install cargo, there is no need to depend
on the standalone cargo package anymore. Instead add host-rustc to the
dependency list.

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Patrick Havelange <patrick.havelange@essensium.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: David Pierret <david.pierret@smile.fr>
Tested-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/rust: drop dependency on host-cargo-bin

As host-rust depends on host-rust-bin, which depends on host-cargo-bin,
remove unneeded dependency.

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Patrick Havelange <patrick.havelange@essensium.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: David Pierret <david.pierret@smile.fr>
Tested-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/rust-bin: host variant depends on host-cargo-bin

Ensure installation of pre-built cargo binary by adding it to the
dependency list of the host variant of the pre-built Rust compiler.

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Patrick Havelange <patrick.havelange@essensium.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: David Pierret <david.pierret@smile.fr>
Tested-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/cargo-bin: install if rust-bin selected

If host-rust-bin, the pre-built Rust compiler, is selected as a rustc
provider, then also install the pre-built cargo binary to be coherent.

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Patrick Havelange <patrick.havelange@essensium.com>
[Arnout: don't install cargo config, alread moved]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: David Pierret <david.pierret@smile.fr>
Tested-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/rust: build and install cargo

Cargo source code is not provided anymore as a separate tarball but is
now built along with the Rust compiler. So update rust host variant to
build Cargo.

For now, all this will be overwritten again by the host-cargo package,
but this package will be removed in later commits.

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Patrick Havelange <patrick.havelange@essensium.com>
Tested-by: Sam Voss <sam.voss@rockwellcollins.com>
[Arnout: merge all install commands in HOST_RUST_INSTALL_CMDS]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/rustc: install cargo config as toolchain post-install hook

The cargo config file is much like the cmake and meson configuration
files, something that tells the cargo build system how the
cross-toolchain is configured.

It is currently installed by the cargo package, but this package is
about to be removed as cargo is now built as part of rust
itself. Therefore, install the cargo config file as a toolchain
post-install hook, just like we do for cmake and meson.

However, we don't have a pkg-cargo infrastructure (yet) so we can't
put it there. Therefore, put it in the rustc package - any cargo
package will need rustc.

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/libnss: bump version to 3.55

Bump version to 3.55. Unfortunately local patch is still pending
upstream and not applied and I've just asked why:
https://bugzilla.mozilla.org/show_bug.cgi?id=1642174

This version requires NSPR 4.27.

Release Notes:
https://developer.mozilla.org/en-US/docs/Mozilla/Projects/NSS/NSS_3.55_release_notes

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
[yann.morin.1998@free.fr: two spaces in hash file]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libnspr: bump version to 4.27

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
[yann.morin.1998@free.fr: two spaces in hash file]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/docker-engine: needs more runtime dependencies

fix error:
failed to start daemon: Devices cgroup isn't mounted

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/mbpfan: new package

Adding mbpfan, a fan control daemon for MacBook laptops.

Signed-off-by: Christian Stewart <christian@paral.in>
[Peter: add hash file, drop Debian init script, drop systemd symlinks, add
 DEVELOPERS entry]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/vlc: fix missing headers when compiling with Qt 5.15

Fixes
http://autobuild.buildroot.net/results/832/832226c43549befda7e8c86a52354e7d6adc0a98/

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/tpm2-tools: bump version to 4.1.2

Bugfix release with a single fix:

- Fix missing handle maps for ESY3 handle breaks. See #1994

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

package/tpm2-abrmd: bump version to 2.3.2

Fixes various initialization / systemd issues. From the changelog:

- Provide meaningful exit codes on initialization failures.
- Prevent systemd from starting the daemon before udev changes ownership of
  the TPM device node.
- Prevent systemd from starting the daemon if there is no TPM device node.
- Prevent systemd from restarting the daemon if it fails.
- Add SELinux policy to allow daemon to resolve names.
- Add SELinux policy boolean (disabled by default) to allow daemon to
  connect to all unreserved ports.

Also adjust .hash file white space to match new agreements.

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

package/libwebsockets: add fsmount support

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libwebsockets: add libglib2 support

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libwebsockets: bump version to 4.0.20

Other changes:
  - Remove upstream patch.
  - Change license type to MIT.
  - Update license shasum due to the above change.
  - Upstrem agained optional support for libglib2 and fsmount, we
    disable those so far

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
[yann.morin.1998@free.fr: forcibly disable new features]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libwebsockets: add support for extensions

libwebsockets currently depends on zlib, however, as of commit
13ba5bbc633ea962d46d9f2bfb2e87c9f7e83b62 (from 2013!) zlib is not actually a
requirement if extensions are not enabled.

By default, libwebsockets sets the "LWS_WITHOUT_EXTENSIONS" option to "ON,"
which removes the need for zlib as a dependency. Because of this default
setting, there are two options:

1) Remove the dependency on zlib, create a new option under Config.in to allow
a user to enable extensions, add zlib as a dependency, and set
LWS_WITHOUT_EXTENSIONS to "OFF" if that option is selected.

2) Keep the dependency and explicitly set LWS_WITHOUT_EXTENSIONS to "OFF."

Below are the size differences in bytes:
  - Without zlib and extensions: 44636
  - With zlib and without extensions: 44720
  - With zlib and with extensions: 44732

Considering the size difference between all 3 is an incredible 96 bytes, and
the difference between what we have now, and enabling extensions is an
insurmountable 12 bytes, it is safe to keep the dependency on zlib and always
enable extensions.

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/powerpc-utils: bump to version 1.3.7

Update indentation in hash file (two spaces)

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/librtas: bump to version 2.0.2

Update indentation in hash file (two spaces)

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/snmppp: fix SNMPv3 with gcc <= 5

Fixes:
 - http://autobuild.buildroot.org/results/bfe3faefc007f2e746c9f0af9b7d0926bb72dc9e

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Reviewed-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/freeswitch: fix build error

Freeswitch has been tested agains tan older version of libcheck, 0.12.0
or earlier, while Buildroot has 0.15.1. With 0.13.0, the API of libcheck
has changed, and Freeswitch does not account for this.

Fixes
http://autobuild.buildroot.net/results/5db/5dbce5860f57c1457962aa295857689f49a60bc6/

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Tested-by: Heiko Thiery <heiko.thiery@gmail.com>
[yann.morin.1998@free.fr: expand commit log with the root cause]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/chocolate-doom: bump to version 3.0.1

Update indentation in hash file (two spaces)

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/freeswitch-mod-bcg729: bump version to 1.0.5

Reformatted hashes.

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

package/libhdhomerun: bump version to 20200521

Reformatted hashes.

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

package/libopenh264: bump version to 2.1.1

Release notes:
https://github.com/cisco/openh264/blob/openh264v2.1.1/RELEASES

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

package/dovecot-pigeonhole: bump version to 0.5.10

Release notes:
https://raw.githubusercontent.com/dovecot/pigeonhole/release-0.5.10/NEWS

Reformatted hashes.

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

boot/barebox: bump version to 2020.07.0

Signed-off-by: Bartosz Bilas <b.bilas@grinn-global.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/check: bump version to 0.15.1

Release notes: https://github.com/libcheck/check/blob/0.15.1/NEWS

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

package/moarvm: bump to 2020.07

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/xapian: bump to version 1.4.16

Updated indentation in hash file (two spaces)

Signed-off-by: Gilles Talis <gilles.talis@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/ncdu: bump to version 1.15.1

Updated indentation in hash file (two spaces)

Signed-off-by: Gilles Talis <gilles.talis@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/libolm: bump to version 3.1.5

Updated hash file indentation (two spaces)

Signed-off-by: Gilles Talis <gilles.talis@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

docs/website: update for 2020.02.4

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

Update for 2020.02.4

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit dee53013da87dfa4bcb3433bdef79ec43b5a5c24)
[Peter: drop Makefile changes]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

docs/website: update for 2020.05.1

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

Update for 2020.05.1

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 667adccb927395ab99d17fc8877b21bf70614218)
[Peter: drop Makefile changes]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/libwebsockets: remove dependency on mmu

Six years ago, commit b6b5bb518d28850e024b90db1129b7f86e9d416c added the MMU
dependency for the libwebsockets package. However, according to the git
history of the CMakeLists.txt file, libwebsockets has supported the vfork
function for at least the last five years.

After testing with the qemu_arm_versatile_nommu_defconfig and the
br-arm-cortex-m4-full.config file, no errors occurred when building
libwebsockets without MMU support.

Remove the dependency; it is no longer necessary. Update the reverse
dependencies as needed.

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/enchant: bump to version 2.2.8

- Update site to get latest version
- Update hash of COPYING.LIB (update in FSF address with
  https://github.com/AbiWord/enchant/commit/adcb08292d4650a1c258f9b02c53cf1e26b47185)
- Update indentation in hash file (two spaces)

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/qlibc: bump to version 2.4.5

Update indentation in hash file (two spaces)

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>