package/qt5base: drop 5.6 support

And get rid of the 5.12.7 subdir now that the version selection is gone.

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

package/qt5: drop 5.6 support

As discussed during the FOSDEM2019 develop days, Qt 5.6 is very old (5.6.3
was released in September 2017, and 5.6.x became EOL in March 2019), so drop
it before the new Buildroot LTS release:

https://elinux.org/Buildroot:DeveloperDaysFOSDEM2019#Qt5_versions_to_support:_keep_5.6_or_a_newer_LTS.3F

And add legacy handling for it.

There are a number of places where code checks for
BR2_PACKAGE_QT5_VERSION_LATEST, so leave that as a blind option for now to
not break the build.

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

package/eudev: support openrc init

Select udev-gentoo-scripts and avoid installing the sysv init script.

Signed-off-by: Carlos Santos <unixmania@gmail.com>
[yann.morin.1998@free.fr: use a fake openrc init hook]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/udev-gentoo-scripts: new package

Gentoo scripts to initialize udev via openrc, to be used by eudev.

Signed-off-by: Carlos Santos <unixmania@gmail.com>
[yann.morin.1998@free.fr: fix commit title]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/python-lmdb: bump to version 0.98

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/{protobuf, python-protobuf}: bump to version 3.11.3

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-pip: bump to version 20.0.2

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-pexpect: bump to version 4.8.0

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-passlib: bump to version 1.7.2

License hash change is due to date update.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-paramiko: bump to version 2.7.1

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python{,3}-mako: bump to version 1.1.1

License hash change is due to date update.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-lxml: bump to version 4.5.0

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-jinja2: bump to version 2.11.1

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-ecdsa: bump to version 0.15

Select new requirement python-six.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-crossbar: bump to version 20.1.2

Rebased avoid intentional syntax error patch.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-asn1crypto: bump to version 1.3.0

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-cffi: bump to version 1.14.0

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-txaio: bump to version 20.1.1

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-autobahn: bump to version 20.2.1

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/musl: bump to version 1.2.0

Also:
 * Drop patch 0003 which is included in the new version.
 * Update hash value of the COPYRIGHT file as the copyright year was
   updated.

>From the release history [1]:

This release moves all 32-bit archs to 64-bit time_t, enabling them
to represent times beyond January of 2038.

There are no new requirements on kernel version, and this is not a
hard ABI break, but the type changes do impact compatibility between
code built against previous versions of musl and code built against
musl 1.2. Users upgrading 32-bit systems should read the detailed
time64 release notes [2]. 64-bit systems are not affected.

In addition, character data has been updated to align with Unicode
12.1.0, along with fixes for some errors in the data and a replacement
for inefficient and unmaintainable case-mapping code. Correctness of
results has been improved in the math library, particularly some complex
functions and 32-bit x86 asm. Various arch-specific bugs have also been
fixed.

[1] https://musl.libc.org/releases.html
[2] https://musl.libc.org/time64.html

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/botan: bump to version 2.13.0

- Calculate hash of tarball locally, hash of 2.13.0 is not in
  https://botan.randombit.net/releases/sha256sums.txt
- Update hash of license file (update in year:
  https://github.com/randombit/botan/commit/3567c921c102448fabc8c2643a1cb6acdb6f9e5c)
- Update indentation of hash file (two spaces)

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

package/scrypt: bump to version 1.3.0

Use the new COPYRIGHT file as the license file

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

package/{avro-c, python-avro}: bump to version 1.9.2

- Drop patch for avro-c that was applied upstream
 - Fix inconsistency in PYTHON_AVRO_SITE
 - Adopt new spacing convention in .hash files
 - Add upstream patch for python-avro
 - Add patch to remove installation-time linting in python-avro

Signed-off-by: Titouan Christophe <titouan.christophe@railnova.eu>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/libmbim: bump to version 1.22.0

https://lists.freedesktop.org/archives/libmbim-devel/2020-January/001102.html

Signed-off-by: Aleksander Morgado <aleksander@aleksander.es>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/openlayers: bump version to 6.2.1

Signed-off-by: Thomas Claveirole <thomas.claveirole@green-communications.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/gnuplot: bump to version 5.2.8

0001-configure-add-without-demo-option: adapt patch to 5.2.8 version
0002-use-gdlib-config-properly: already applied upstream

Signed-off-by: Michael Fischer <mf@go-sys.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/syslog-ng: bump version to 3.25.1

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/linux-firmware: bump to version 20200122

Added new r8169 firmware files.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-flatbuffers: new package

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/fswebcam: bump to latest git version

Bump to latest git version.

Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-markdown: bump to version 3.2.1

Update indentation of hash file (two spaces)

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/libxkbcommon: bump to version 0.10.0 and convert to meson

Since version 0.9.x the autotools build system is removed from libxkbcommon
([1]), so convert to meson build system.

Remove the CFLAGS force of '-std=gnu99', believing meson sets it
automatically (a test build used '-std=c99').

For details see [1].

[1] https://lists.freedesktop.org/archives/wayland-devel/2019-October/040962.html

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Reviewed-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

utils/scanpypi: update hash file indentation formatting

The new .hash convention is to use 2 spaces between fields.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/vala: bump version to 0.46.6

Also add the line "HOST_VALA_CONF_OPTS += --disable-valadoc"
as valadoc requires libgvc which is not available in Buildroot.

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

package/kvm-unit-tests: drop x86_64 workaround

This workaround is not needed since commit
0788e921f961e36a50a94f8bb98f7eb75b312245 which disables stack-protector
for all architectures as intended by upstream

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

package/kvm-unit-tests: fix build with SSP

Add a patch to correct a typo in the Makefile, so -fno-stack-protector /
-fno-stack-protector-all are really used.  With this applied, kvm-unit-tests
will always be built without SSP as intented by upstream.  This will fix the
build on ppc64 with SSP that started to fail for an unknown reason since
November 27th.

Moreover, the Arch Linux workaround could also be removed in a follow-up
patch.

Fixes:
 - http://autobuild.buildroot.org/results/ad689b08173548af21dd1fb0e827fd561de6dfef

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

package/python-msgpack: bump to version 1.0.0

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/psplash: bump to revision fd33a9b

This includes the following changes:

66c8f65 Makefile.am: use BUILT_SOURCES instead of BUILD_SOURCES
f7a354d Generate image headers from their PNG source, while still including them in dist tarball
983837e Import png source for images, extracted from headers
3bd1557 Cleanup poky logo
773a397 Add configure flags for fullscreen logo and no startup message
2913c24 Double buffering support
8188d68 Support --with-font configure option to use other font headers
88c4b71 Add SPDX License information to files, drop boilerplate
904f4aa add systemd support
e94bd11 process consecutive commands
568f28e use /run for communication FIFO
ee053e0 add recovered bar.png
2015f70 Fix text width calculation.

Also:
 - remove the 'psplash-quit.service' unit file
 - add 'psplash-systemd.service' to use the new systemd support in psplash

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

package/libsecret: bump to version 0.20.1

- Drop patch (already in version)
- Update indentation of hash file (two spaces)

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

package/meson: bump version to 0.53.2

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

package/armadillo: bump to version 9.850.1

Update indentation of hash file (two spaces)

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

package/openblas: bump to version 0.3.8

Update indentation of hash file (two spaces)

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

package/i2pd: bump to version 2.30.0

Update indentation of hash file (two spaces)

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

package/libcpprestsdk: bump to version 2.10.15

Update indentation of hash file (two spaces)

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

package/drbd-utils: bump to version 9.12.0

Update indentation of hash file (two spaces)

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

package/pcre2: bump to version 10.34

Update indentation of hash file (two spaces)

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

package/pcre: bump to version 8.44

- Update first patch
- Update hash of license file (update in year)
- Update indentation of hash file (two spaces)

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

package/sed: bump to version 4.8

Update indentation of hash file (two spaces)

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

package/upmpdcli: bump to version 1.4.6

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

support/scripts/pkg-stats: iterate over CVEs in streaming

The NVD files that are used to build the list of CVEs affecting
Buildroot packages are quite large (a few hundreds MB of json),
and cause the pkg-stats scripts to have a huge memory footprint
(a few GB with Python 2.7).

However, because we only need to iterate on CVE items one by one,
we can process them in streaming (ie decoding one CVE at a time
from the JSON representation). Because the json module from the
python standard library does not support such a mode of operation,
we switch to the third-party package ijson, which is compatible
with both Python 2 and Python3.

To run the script with these modifications, one should install
the ijson python package. This can be done with pip:
`pip install ijson`. On Debian based distributions, this can
also be done with the apt package manager:
`apt install python-ijson`.

Signed-off-by: Titouan Christophe <titouan.christophe@railnova.eu>
Reviewed-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Tested-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/kmod: bump to version 27

- Drop second and third patches (already in version)
- Update indentation of hash file (two spaces)

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

package/avahi: bump to version 0.8

This bump also includes:
 * Drop upstream security patch which is included in the new version
 * Unconditionally disable support for Qt5 [1] (same as Qt3 and Qt4)
 * Drop dependency on host-inttool, as avahi switched to host-gettext [2]
 * Conditionally enable support for libevent [3]

[1] https://github.com/lathiat/avahi/commit/5dbb32767ae3f5a371cfbd04b4e3a9a634b8efc4
[2] https://github.com/lathiat/avahi/commit/3d5a0c68057e2ed76187a0bb565baaa10d566003
[3] https://github.com/lathiat/avahi/commit/998e20cd76927ce978fb5676820a38308e21f45d

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/kodi-visualisation-fishbmc: remove opengl dependency

The addon can also be used with OpenGLES:
https://github.com/xbmc/visualization.fishbmc/blob/Leia/CMakeLists.txt

Updated project URL forgotten in
https://git.buildroot.net/buildroot/commit/?id=4726e93c63d51fe287c62628a9c71c04517d727c

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

package/sispmctl: bump to version 4.2

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

package/gobject-introspection: add missing newline at end of wrappers

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/smartmontools: fix build without stack-protector

Fixes:
 - http://autobuild.buildroot.org/results/0de9f2a69fa2a39164211299f8a429d2fec6935a

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

package/{mesa3d, mesa3d-headers}: bump version to 20.0.0

Added optional dependency to zstd present due to upstream commit:
https://cgit.freedesktop.org/mesa/mesa/commit/?id=a8d941091f72923561a6c58b46ccb264b6a0e205

Replaced MESA_EGL_NO_X11_HEADERS define with EGL_NO_X11 due to upstream
commit:
https://cgit.freedesktop.org/mesa/mesa/commit/?id=6202a13b71e18dc31ba7e2f4ea915b67eacc1ddb

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

package/smartmontools: bump to version 7.1

- systemd optional dependency has been added in version 7.0
- Update indentation of 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/gobject-introspection: depend on python3

Currently, the Config.in file has the line:
select BR2_PACKAGE_PYTHON3 if !BR2_PACKAGE_PYTHON

This line is incorrect as gobject-introspection does not support python2.
Instead, remove the select line and make python3 a dependency with a new
message that explains that gobject-introspection requires python3.

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
[yann.morin.1998@free.fr:
  - move the explanations from the commit log to the code
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libeXosip2: bump to version 5.1.1

- Removed patch that was applied upstream
- Removed AUTORECONF request that came with patch
- Updated library's download name

Signed-off-by: Gilles Talis <gilles.talis@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libosip2: bump to version 5.1.1

Removed patch that was applied upstream

Signed-off-by: Gilles Talis <gilles.talis@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/gobject-introspection: new package

GObject introspection is a middleware layer between C
libraries (using GObject) and language bindings. The C library
can be scanned at compile time and generate a metadata file,
in addition to the actual native C library. Then at runtime,
language bindings can read this metadata and automatically
provide bindings to call into the C library.

There's an XML format called GIR used by GObject-Introspection.
The purpose of it is to provide a standard structure to access the complete
available API that a library or other unit of code exports. It's
language-agnostic using namespaces to separate core, language, or
library-specific functionality.

Cross-compiling gobject-introspection is not an easy task. The main issue is
that in the process of creating the XML files, gobject-introspection must first
run and scan the binary, which, if the binary is cross-compiled, would not
typically be possible from the host system.

Because of this limitation, we use several wrappers to call instead first out
qemu, which runs the native scanner to create the binaries.

There are seven total patches and four different wrapper files needed to
successfully cross-compile and run this package, many of them are from
open-embedded, but one of them is of my own doing.

1) Revert a previous, incomplete attempt at adding cross-compiling support.

2) Add support for cross-compiling with meson.

3) Disable tests.

4) Add an option to use a binary wrapper; this patch will force giscanner to
   use a wrapper executable to run binaries it's producing, instead of
   attempting to run them from the host.

5) Add an option to use an LDD wrapper, again, useful for cross-compiled
   environments.

6) Add a --lib-dirs-envar option to pass to giscanner. (See patch for details.)

7) Add rpath-links to ccompiler: when passing the PACKAGE_GIR_EXTRA_LIBS_PATH
   to the ccompiler.py script, ccompiler.py needs to add -Wl,-rpath-link to the
   environment for the package to correctly link against the passed on paths.

8) Ignore error return codes from ldd-wrapper because prelink-rtld returns 127
   when it can't find a library, which breaks subprocess.check_output().

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
Tested-by: Yegor Yefremov <yegorslists@googlemail.com>
Tested-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
[yann.morin.1998@free.fr:
  - host-prelink-cross has no Kconfig entry
  - reorder dependencies for arch deps first
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/pkgconf: prepend sysroot paths to common gobject-introspection utils

Many autotools packages call pkg-conf to inquire as to where the following
utilities are:

g_ir_scanner
g_ir_compiler
g_ir_generate

Because gobject uses wrappers to call qemu, prepending the sysroot to the paths
of these compilers is necessary.

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

package/prelink-cross: new package

Prelink-cross emulates a runtime linker for a given sysroot. This is
necessary to allow gobject-introspection to build its typelib files
during cross-compiling.

We're using a sha1 on the cross_prelink branch, as we need the
RTLD-enabled variant of prelink-cross.

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
[yann.morin.1998@free.fr:
  - drop HOST_ prefix for inherited variables
  - fix licensing info to "or-later"
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/gdb: depend on libiberty

If present, GDB may use a system installed libiberty. As such, we must ensure
that host-libiberty is installed first.

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

package/libiberty: new package

Some packages, like prelink-cross, want to use libiberty but do not bundle
their own instance (which is good!).

However, libiberty is made for being bundled in packages: all GNU
packages that use libiberty (gcc, Binutils, gdb, et al...) all have their own
bundled variant. This common practice means that there is no official upstream
for libiberty, the closest being as part of the combined Binutils-gdb tree.

So we introduce a new host-only package, that installs just libiberty from a
Binutils released tarball.

Again, as packages usually bundle libiberty, it usually only installs a static
version. Furthermore, it does not obey the usual --enable-shared and
--disable-static flags; it only ever builds a static version.

Furthermore, -fPIC is not used with this library, but some packages may pick it
to build shared objects. This behavior is the case for host-gdb, for example,
which accidentally picks that library instead of its internal one.

So, rather than fix the various gdb versions and variants we can use, we ensure
that the libiberty we install is usable in shared objects, and we always build
before host-gdb.

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
[yann.morin.1998@free.fr:
  - fix DL_SUBDIR for a host-only package
  - add licensing info
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/gensio: bump to version 1.5.2

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/mpd: bump version to 0.21.20

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/openfpgaloader: new package

openFPGALoader is a tool for programming FPGA.

Signed-off-by: Jean Burgat <jeanburgat33@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/avahi: drop most of the AVAHI_CONF_ENV vars

Most are legacy from when the package was added and not really necessary.

This commit is based on dropping the CONF_ENV vars in libgtk2 [1].

[1] https://git.buildroot.net/buildroot/commit/package/libgtk2/libgtk2.mk?id=4d80cbdf6c2c186da26f36575d6635604f00d29a

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/libubootenv: bump to version ba952d0

This includes the following changes:

ba952d0 BUG: variable lists not released in close()
690f868 Variables are not removed when loading from file
9e3586a Make sure there's no file descriptor leakage in case of error
03647c4 Check config file defines a non-zero Sector size
3b2d4f1 Check environment size from fw_env.config

Signed-off-by: Pierre-Jean Texier <pjtexier@koncepto.io>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/uacme: bump version to 1.0.22

Signed-off-by: Nicola Di Lieto <nicola.dilieto@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/libiio: bump to version 0.19

See full changelog https://github.com/analogdevicesinc/libiio/releases/tag/v0.19

Also remove patch applied upstream

Signed-off-by: Pierre-Jean Texier <pjtexier@koncepto.io>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

Update for 2020.02-rc1

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

CHANGES: update with recent changes

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

package/libsigrok: explain why host-doxygen is needed

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

package/owfs: fixup Python sysconfigdata for per-package directories

This is needed so that building the owfs Python module uses the gcc
from owfs per-package directory, and not the one from the python
per-package directory.

Fixes:

  http://autobuild.buildroot.net/results/0d582dda367507991a4c38141db36b0fa8e47e67/

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

package/pkg-python: fix for per-package directories

With per-package directory support, Python external modules are
causing a problem: the _sysconfigdata.py module installed by the
Python interpreter contains a number of paths that are relative to the
current package per-package directory, i.e python or python3. For
example:

'BLDSHARED': '/home/thomas/projets/buildroot/output/per-package/python/host/bin/arm-linux-gcc -shared',
'CC': '/home/thomas/projets/buildroot/output/per-package/python/host/bin/arm-linux-gcc',
'CXX': '/home/thomas/projets/buildroot/output/per-package/python/host/bin/arm-linux-g++',
etc.

These paths are problematic, because it means that the wrong compiler
gets used when building external Python modules: instead of using the
compiler from the external Python module per-package host directory,
it uses the one from the 'python' or 'python3' per-package host
directory. Due to this, any native dependency needed by the external
Python module is not found, even though it is properly present in the
current package per-package directory.

Of course, the problem occurs with both target Python modules and host
Python modules.

To fix this, we simply rewrite those paths in _sysconfigdata.py before
building a Python package.

Interestingly, until now, the _sysconfidata.py that was used during
the build was the one from $(TARGET_DIR), which is a bit unusual: it
is more common to use files from $(STAGING_DIR) during the build
process. So this commit changes the PYTHON_PATH and PYTHON3_PATH
variables so that they point to $(STAGING_DIR), which makes the
_sysconfigdata.py fixup in $(STAGING_DIR) effective.

Fixes:

  http://autobuild.buildroot.net/results/a24b0555fd4261b50dc3986635c30717d9cbe764/ (python-psycopg2)
  http://autobuild.buildroot.net/results/080fa893e1b0e7a8c8a31ac1c98eb8871b97264d/ (python-alsaaudio)
  http://autobuild.buildroot.net/results/79bc070f98d6d9d8ef78df12b248cdc7d0e405c3/ (python-lxml)
  and many more Python packages that use native code with a native library

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

package/apache: fix build with per-package directory support

When APR_INCLUDEDIR and APU_INCLUDEDIR point to the same directory,
Apache builds properly. However, with per-package directory support,
they point to different directories, and APU_INCLUDEDIR contains both
the APR headers and the APU headers.

Due to this, the Apache Makefile logic to generate its exports.c file
leads to duplicate definitions, because the APR headers are considered
twice: once from APR_INCLUDEDIR, once from APU_INCLUDEDIR.

We fix this by introducing a patch to the Apache build system.

In addition, apr provides a special libtool script that gets used by
apr-util and apache. apr-util already had a fixup for this, but apache
did not, which was causing the gcc from apr-util per-package
directories be used during the apache build, causing build failures.

To fix this, we adjust this libtool script to point to the correct
tools in apache's per-package directories.

There are no autobuilder failures for this, because Apache needs
apr-util, and apr-util currently fails to build when
BR2_PER_PACKAGE_DIRECTORIES=y.

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

package/apr-util: fix build with per-package directories

With per-package directories support enabled, the build of apr-util
fails, for two reasons:

 - The rules.mk file is generated by the 'apr' package, and then
   copied into the 'apr-util' source directory. This is done by the
   'apr-util' build process. Unfortunately, this rules.mk file has a
   number of hardcoded paths: to the compiler and to the libtool
   script.

   Due to this, the compiler from the 'apr' per-package directory gets
   used. But this compiler uses the 'apr' package sysroot, which does
   not have all the dependencies of the 'apr-util' package, causing
   the build to fail because <expat.h> is not found.

 - Similarly, the libtool script itself has some hardcoded paths,
   which make it use the compiler/linker from the 'apr' per-package
   directory, so it does not find the expat library.

We fix both issues by doing the necessary replacement in both rules.mk
and libtool.

Fixes:

  http://autobuild.buildroot.net/results/2a67b5d58f79348e20a972125e4797eff5585716/

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

package/cog: add patch fixing cog segfault

Fixes:
Thread 1 "cog" received signal SIGSEGV, Segmentation fault.
xkb_state_update_mask (state=0x0, base_mods=0, latched_mods=0, locked_mods=0, base_group=base_group@entry=0, latched_group=latched_group@entry=0, locked_group=0) at ../src/state.c:814
814     prev_components = state->components;

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Acked-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/libxml2: add upstream security fix for CVE-2019-20388

Fixes CVE-2019-20388: xmlSchemaPreRun in xmlschemas.c in libxml2 2.9.10
allows an xmlSchemaValidateStream memory leak.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/pulseview: depends on host gcc >= 4.9

Commit 88bb278d5ac790bee0c3a438464da82ee7625cff forgot to propagate the
new host gcc >= 4.9 dependency from BR2_PACKAGE_LIBSIGROKCXX

Fixes:
 - http://autobuild.buildroot.org/results/5dc9dc95d0534b35e2443c120162b5176edafe0b

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

package/nodejs: security bump to version 12.16.0

Fixes the following security issues (12.15.0):

- CVE-2019-15606: HTTP header values do not have trailing OWS trimmed

- CVE-2019-15605: HTTP request smuggling using malformed Transfer-Encoding
  header

- CVE-2019-15604: Remotely trigger an assertion on a TLS server with a
  malformed certificate string

For more details, see the advisory:
https://nodejs.org/en/blog/vulnerability/february-2020-security-releases/

On top of this, 12.16.0 brings a number of changes and bugfixes.

Update the license hash for an addition of the (MIT) licensing terms for the
uvwsai module:

+
+- uvwasi, located at deps/uvwasi, is licensed as follows:
+  """
+    MIT License
+
+    Copyright (c) 2019 Colin Ihrig and Contributors
+
+    Permission is hereby granted, free of charge, to any person obtaining a copy
+    of this software and associated documentation files (the "Software"), to deal
+    in the Software without restriction, including without limitation the rights
+    to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+    copies of the Software, and to permit persons to whom the Software is
+    furnished to do so, subject to the following conditions:
+
+    The above copyright notice and this permission notice shall be included in all
+    copies or substantial portions of the Software.
+
+    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+    IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+    FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+    AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+    LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+    OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+    SOFTWARE.
+  """

While we are at it, adjust the white space in the .hash function to match
the new agreements.

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

package/qpdf: fix build with gcc 4.8

Fixes:
 - http://autobuild.buildroot.org/results/ad7fb68ae87850a85509eed80fd0cae8721b10c5

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

package/gutenprint: add back the hook for creating the m4local directory

Commit 64c42c5e2c26261e26c3548c86b02f55d12f341b removed the hook for
creating the m4local directory with the assumption that it would be
created because the first include is treated in a special way if it
doesn't exists

However, this assumption was wrong as m4local is the second include, the
first one is m4 (which already exists in the archive). So put back the
hook. The other solutions would be to patch:
 - Makefile.{am,in} to remove m4local
 - configure.ac and Makefile.{am,in} to add m4local before m4
However, both solutions don't seem to be upstreamable

Fixes:
 - http://autobuild.buildroot.org/results/e40313c6ec193d6156e26eff62303545fba09413

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

core: fix packages-file-list.txt after an incremental build

The package instrumentation step 'step_pkg_size' is populating the files:
    output/build/packages-file-list.txt
    output/build/packages-file-list-staging.txt
    output/build/packages-file-list-host.txt
by comparing the list of files before and after installation of a package,
with some clever tricks to detect changes to existing files etc.

As an optimization, instead of gathering this list before and after each
package, where the 'after-state' of one package is the same as the
'before-state' of the next package, only the 'after-state' is used and
is shared between packages.

This works fine, except at the end of the build, as explained next.

In the target-finalize step, many files will be touched. For example, files
like /etc/hosts, /etc/os-release, but also all object files that are
stripped, and all files touched by post-build scripts or created by rootfs
overlays. This means that the 'after-state' of the last package does not
reflect the actual situation after target-finalize is run.

For a single complete build this poses no problem. But, if one incrementally
rebuilds a package after the initial build, e.g. with 'make foo-rebuild',
then all changes that happened in target-finalize at the end of the initial
build (the 'after-state' of the last package built) will be detected as
changes caused by the rebuild of package foo. As a result, all these files
will incorrectly be treated as 'owned' by package foo.

Correct this situation by capturing a new state at the end of
target-finalize, so that the 'before-state' of an incremental build will be
correct.

Note: the reasoning above talks about packages-file-list.txt and
target-finalize, but also applies to
packages-file-list-staging.txt/staging-finalize and
packages-file-list-host.txt/host-finalize.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

support/run-tests: reorder imports

Reorder imports using the isort utility to fix a warning from pylint3:

wrong-import-order: standard import "import multiprocessing" should be
placed before "import nose2"

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

package.nfs-utils: drop extra empty line

Commit 12c0f68caf (package/nfs-utils: bump version to 2.4.3) added an
extra empty line, causing check-package to whine:

    package/nfs-utils/nfs-utils.mk:27: consecutive empty lines

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

configs/qemu{x86, x86_64}: add a serial console

The current Buildroot defconfigs for qemu_x86 and qemu_x86_64
instantiate a console on tty1, which appears on QEMU's
graphical window. Add a console on the serial port (ttyS0) to
be used later for gitlab testing.

This change is need since the script used for gitlab testing
needs to use a serial output with pexpect.

This change is similar to the one made for raspberrypi [1] to
handle HDMI and serial console:

This requires three changes:
 1. have two 'console=' entries in the kernel command line: tty1,
    then ttyS0;
 2. change BR2_TARGET_GENERIC_GETTY_PORT to "console", so it starts
    a getty on the last console= passed to the kernel, ttyS0;
 3. add a new getty on tty1 to the generated inittab.

Step 2 is actually obtained by removing BR2_TARGET_GENERIC_GETTY_PORT
entirely from the defconfigs, since "console" is the default value.

Step 3 requires a post-build script since the Buildroot makefiles can
configure only one console.

Note: instead of simply adding a new getty on ttyS0 (which would
work) this patch actually changes BR2_TARGET_GENERIC_GETTY_PORT to
instantiate a console on UART, then adds back tty1 via
post-build.sh. This is done only to avoid the "GENERIC_SERIAL" comment
where we instantiate a console on QEMU graphical window, then
instantiate a really-serial console on another line.

The result is these two inittab lines:

  console::respawn:/sbin/getty -L  console 0 vt100 # GENERIC_SERIAL
  tty1::respawn:/sbin/getty -L  tty1 0 vt100 # QEMU graphical window

[1] 20878a1017e2bf7eb8c5f870dc6d2641493cb0f9

Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/qemu_pcc_mac99: build host-qemu for runtime testing

The commit [1] added host-qemu package for each qemu defconfig
for gitlab runtime testing.

[1] 29e1cb88844614c40846540e22cf83aa9e52674f

Signed-off-by: Romain Naour <romain.naour@smile.fr>
Cc: Joel Stanley <joel@jms.id.au>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/qemu_ppc_mac99_defconfig: add usual comments for Kconfig symbols

This defconfig was generated by savedefconfig but we usually
use a manually modified defconfig to add some comments for
Kconfig symbols.

No content change intended.

Signed-off-by: Romain Naour <romain.naour@smile.fr>
Cc: Joel Stanley <joel@jms.id.au>
Acked-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/janus-gateway: bump version to 0.8.1

Other changes:
  - Update License hash which properly adds the OpenSSL exception.

Tested with Debian 8:

br-arm-full [1/6]: OK
br-arm-cortex-a9-glibc [2/6]:   OK
 br-arm-cortex-m4-full [3/6]:   SKIPPED
        br-x86-64-musl [4/6]:   OK
    br-arm-full-static [5/6]:   SKIPPED
          sourcery-arm [6/6]:   OK

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

package/qemu: Bump to version 4.2.0

Other changes:
  - Remove upstream patches
  - Update COPYING.LIB hash as upstream updated the file to match the new LGPL
    2.1 license from upstream. See:
    https://github.com/qemu/qemu/commit/f0d44cc4462f112bce5ec556e87eff4eec682e39

Signed-off-by: Adam Duskett <Aduskett@gmail.com>
Tested-by: Romain Naour <romain.naour@gmail.com>
[Peter: change libssh2 to libssh as pointed out by Vincent Fazio]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/nfs-utils: bump version to 2.4.3

Bump to version 2.4.3 of nfs-utils.  All patches have been upstreamed, so
drop them all.  It now needs rpcgen built by host-nfs-utils, to do this
let's pass its path to --with-rpcgen= instead of 'internal'.

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Reviewed-by: Petr Vorel <petr.vorel@gmail.com>
Tested-by: Petr Vorel <petr.vorel@gmail.com>
[Peter: drop AUTORECONF, explicitly depend on host-nfs-utils]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/minicom: bump version

For a minor fix.

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/glslsandbox-player: remove 'v' prefix

Fixes version parsing for release-monitoring.org support

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

support/run-tests: check for empty sequences in a pythonic way

According to PEP8 empty sequences should be checked as booleans.

Fixes the following PEP8 warning:
Do not use `len(SEQUENCE)` to determine if a sequence is empty

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

{linux, linux-headers}: bump 4.{4, 9, 14, 19}.x / 5.4.x series

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

linux: use correct conditional for wireguard kernel config fixup

Commit de591c5c3a93 (package/wireguard-linux-compat: new package) split up
the wireguard package in wireguard-tools and wireguard-linux-compat, but
forgot to update the conditional in linux.mk, so the kernel config fixups
needed for wireguard are no longer applied.

Update the conditional to use the BR2_PACKAGE_WIREGUARD_LINUX_COMPAT symbol
instead.

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