package/python-boto3: new package.

The AWS SDK for Python.

Signed-off-by: Raphaël Mélotte <raphael.melotte@essensium.com>
Reviewed-by: Yegor Yefremov <yegorslists@googlemail.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Tested-by: Jagan Teki <jagan@amarulasolutions.com> # roc-rk3399-pc
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/python-s3transfer: new package.

An Amazon S3 Transfer Manager.

Signed-off-by: Raphaël Mélotte <raphael.melotte@essensium.com>
Reviewed-by: Yegor Yefremov <yegorslists@googlemail.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Tested-by: Jagan Teki <jagan@amarulasolutions.com> # roc-rk3399-pc
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/python-botocore: new package.

Low-level, data-driven core of boto 3.

Signed-off-by: Raphaël Mélotte <raphael.melotte@essensium.com>
Reviewed-by: Yegor Yefremov <yegorslists@googlemail.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Tested-by: Jagan Teki <jagan@amarulasolutions.com> # roc-rk3399-pc
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/python-jmespath: new package.

JMESPath is a query language for JSON.

Signed-off-by: Raphaël Mélotte <raphael.melotte@essensium.com>
Reviewed-by: Yegor Yefremov <yegorslists@googlemail.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Tested-by: Jagan Teki <jagan@amarulasolutions.com> # roc-rk3399-pc
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

DEVELOPERS: fix alphabetical ordering

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

package/micropython: fix license hash

Commit 7676191cfacbdb1a92af1d0218d3280207b03090 forgot to update
license hash which changed due to an update in year:
https://github.com/micropython/micropython/commit/338b12d3c840ea56504c83d9c51e4d0f628684f4

While at it, also update indentation in hash file (two spaces)

Fixes:
 - http://autobuild.buildroot.org/results/2a650636965d8575374333f847acc169caf8d2ac

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

package/gstreamer1/gst1-plugins-base: gl needs api, platform and window

Build will fail if gl is enabled without an api, a platform and a
window:

Message: No OpenGL Platforms found or requested
Message: No OpenGL Window systems found or requested

gst-libs/gst/gl/meson.build:948:2: ERROR: Problem encountered: GStreamer OpenGL integration required via options, but needed dependencies not found.

This requirement is already specified in Config.in through a simple
comment:

comment "The opengl library needs an API, a platform and a window system"
depends on !BR2_PACKAGE_GST1_PLUGINS_BASE_HAS_LIB_OPENGL

Fixes:
 - http://autobuild.buildroot.org/results/d171059801adf8dea1a2116d7c729a2aa5767ac8
 - http://autobuild.buildroot.org/results/b1bfa505bba534440d331c4948eea5eb2d165c97

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

package/altera-stapl: new package

Alteras Jam STAPL Bytecode Player with 64-bit support.
This is a user-space port of the altera-stapl driver from the linux kernel.
The source released by Altera wasn't 64-bit compatible.
Additionally, the linux version is much cleaner and old cruft was removed.

Signed-off-by: Jochen Baltes <jochen.baltes@gmail.com>
Tested-by: Heiko Thiery <heiko.thiery@gmail.com>
[yann.morin.1998@free.fr: use the latest HEAD of the repository]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/asn1c: new host package

The asn1c is a free, open source compiler of ASN.1 specifications into C
source code.

Signed-off-by: Sergio Prado <sergio.prado@e-labworks.com>
Reviewed-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/pkg-meson: ensure the global cross-compilation.conf file is correct

Currently, the cross-compilation.conf installed in
$(HOST_DIR)/etc/meson/cross-compilation.conf for use by the SDK is
generated in a post-install-staging hook of the toolchain package.

With per-package directory support enabled, this means that the
generated cross-compilation.conf contains references to the
per-package directory of the toolchain/ package, which is not want we
want:

[binaries]
c = '/home/thomas/projets/buildroot/output/per-package/toolchain/host/bin/arm-linux-gcc'
cpp = '/home/thomas/projets/buildroot/output/per-package/toolchain/host/bin/arm-linux-g++'
ar = '/home/thomas/projets/buildroot/output/per-package/toolchain/host/bin/arm-linux-ar'
strip = '/home/thomas/projets/buildroot/output/per-package/toolchain/host/bin/arm-linux-strip'
pkgconfig = '/home/thomas/projets/buildroot/output/per-package/toolchain/host/usr/bin/pkg-config'

So instead, we generate this file in TOOLCHAIN_TARGET_FINALIZE_HOOKS,
so that the global paths are used:

[binaries]
c = '/home/thomas/projets/buildroot/output/host/bin/arm-linux-gcc'
cpp = '/home/thomas/projets/buildroot/output/host/bin/arm-linux-g++'
ar = '/home/thomas/projets/buildroot/output/host/bin/arm-linux-ar'
strip = '/home/thomas/projets/buildroot/output/host/bin/arm-linux-strip'
pkgconfig = '/home/thomas/projets/buildroot/output/host/usr/bin/pkg-config'

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/gstreamer1/gst1-shark: bump version to v0.7.1

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

package/micropython: bump version to 1.13

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

package/syslog-ng: bump to 3.29.1

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

package/systemd: bump version to 246.5

Signed-off-by: Norbert Lange <nolange79@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/f2fs-tools: bump version to 1.14.0

Remove upstreamed patches.

Signed-off-by: Norbert Lange <nolange79@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/ipmitool: fix CVE-2020-5208

Add several upstream patches that are made to fix this CVE. Since there
is still no dated plan to release a new version add this bunch of
patches.

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

package/localedef: depend host build on python3

Just like glibc, host-localedef needs python3 on the host to
build... since host-localedef is basically using the sources of glibc.

Fixes:

checking if /build/build/per-package/host-localedef/host/bin/ccache
/usr/bin/gcc is sufficient to build libc... yes
checking for x86_64-pc-linux-gnu-nm... /usr/bin/nm
checking for python3... no
checking for python... python
checking version of python... 2.7.18, bad
configure: error:
*** These critical programs are missing or too old: python
*** Check the INSTALL file for required versions.

As reported at:

  http://lists.busybox.net/pipermail/buildroot/2020-September/291929.html

Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/mutt: bump to version 1.14.7

This is a bug-fix release, fixing a variety of small issues.

https://gitlab.com/muttmua/mutt/raw/stable/ChangeLog

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

package/libpfm4: bump to version 4.11.0

Major updates:
 - AMD Zen1 core PMU updates
 - AMD Zen2 core PMU support
 - Intel Icelake core PMU support
 - support to report if events are speculative
 - support for Extended PEBS events on Icelake
 - Intel Tremont core PMU support
 - ARM Neoverse N1 core PMU support
 - s390 event table updates
 - Marvell TX2 core and uncore updates
 - Intel CascadeLakeX uncore PMU support
 - Various other event table updates and bug fixes

http://perfmon2.sourceforge.net/

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

package/libupnpp: bump to version 0.19.4

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

package/libnpupnp: bump to version 4.0.11

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

package/libuhttpd: fix tarball name in hash file

Commit 819a4207f6a5c3b579cb5191b03b51ebec4298ff forgot to rename
libuhttpd-3.2.0.tar.gz

Fixes:
 - http://autobuild.buildroot.org/results/965be1c72996df94668be81d1e90ed6937fb9a7c

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

package/pdmenu: bump to version 1.3.6

- Update site to get latest release
- Update indentation in hash file
- This bump will also fix a build failure with gcc 10 thanks to
  https://git.joeyh.name/index.cgi/pdmenu.git/commit/?id=1a2dce9aaa6cdd0406f0be666929562d3b4521ff

Fixes:
 - http://autobuild.buildroot.org/results/78946a381f9f70975b41d24b0b050ba2073d6ef8

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

package/libuhttpd: bump version to 3.3.0

Signed-off-by: Jianhui Zhao <zhaojh329@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/librsvg: fix license

Commit df20e4546384d1c249a6a0541183590671e83d77 forgot to update hash of
COPYING.LIB which has been updated to use LGPL-2.1+ in version 2.47.0:
https://github.com/GNOME/librsvg/commit/96b34f865dc8465d2ab93c7ae8344178b0ea7ed7

Fixes:
 - http://autobuild.buildroot.org/results/94fde5707a1104548c3386b628c92ebc9e034285

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

package/docker-engine: bump to version 19.03.12

Signed-off-by: Christian Stewart <christian@paral.in>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/docker-cli: bump to version 19.03.12

Signed-off-by: Christian Stewart <christian@paral.in>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/docker-containerd: bump to version 1.4.0

Signed-off-by: Christian Stewart <christian@paral.in>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/runc: bump to version 1.0.0-rc92

Signed-off-by: Christian Stewart <christian@paral.in>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/efl: depends on gcc >= 4.9

efl depends on C++11 since version 1.22.0 and
https://github.com/Enlightenment/efl/commit/ac95f38d1b06ac6977ea6d8b9f14f9405396858f

Fixes:
 - http://autobuild.buildroot.org/results/458ec719a7251b59796edfc865b7226baaed6d8f

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

package/sofia-sip: fix build without openssl

HTTP-related modules are enabled by default and depend on openssl since
https://github.com/freeswitch/sofia-sip/commit/5b4103a12596ea01b4cdbb21b75826a085d557aa

Fixes:
 - http://autobuild.buildroot.org/results/95b6c489fed07d6fe626200ea057c83756d54deb

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

package/sofia-sip: fix static build with openssl

Use --with-openssl=pkg-config otherwise sofia-sip will fail to retrieve
openssl dependencies resulting in the following build failure:

/srv/storage/autobuild/run/instance-3/output-1/host/opt/ext-toolchain/bin/../lib/gcc/arm-buildroot-linux-uclibcgnueabi/8.3.0/../../../../arm-buildroot-linux-uclibcgnueabi/bin/ld: ../libsofia-sip-ua/.libs/libsofia-sip-ua.a(tport_type_ws.o): in function `tport_ws_deinit_primary':
tport_type_ws.c:(.text+0x2c): undefined reference to `SSL_CTX_free'
/srv/storage/autobuild/run/instance-3/output-1/host/opt/ext-toolchain/bin/../lib/gcc/arm-buildroot-linux-uclibcgnueabi/8.3.0/../../../../arm-buildroot-linux-uclibcgnueabi/bin/ld: ../libsofia-sip-ua/.libs/libsofia-sip-ua.a(tport_type_ws.o): in function `tport_ws_init_primary_secure':

pkg-config option has been added to freeswitch sofia-sip since 2007 and
https://github.com/freeswitch/sofia-sip/commit/044ac9d8fa7e840cf3e38c3fa7f8cacd0d21b3ea

Fixes:
 - http://autobuild.buildroot.org/results/8a0c9dbad74a3b358f90c7a3b56b638c4bdb8045

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

package/kexec: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/89b6337db9efc55aaa422fc9964f942b6d4cd5f7

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

package/libcap: needs MMU

libcap uses fork() since version 2.33 and
https://git.kernel.org/pub/scm/linux/kernel/git/morgan/libcap.git/commit/?id=ef485973d5547431782f9e5f4323eabfebb38622

So add MMU dependency to avoid the following build failure on
applications using libcap without MMU such as ntp:

/tmp/instance-0/output-1/host/opt/ext-toolchain/arm-buildroot-uclinux-uclibcgnueabi/bin/ld.real: /tmp/instance-0/output-1/host/bin/../arm-buildroot-uclinux-uclibcgnueabi/sysroot/usr/lib/libcap.a(cap_proc.o): in function `cap_launch':
/tmp/instance-0/output-1/build/libcap-2.42/libcap/cap_proc.c:912: undefined reference to `fork'

All packages that select BR2_PACKAGE_LIBCAP already have a BR2_USE_MMU
dependency, so none of them have to be changed.

Fixes:

 - http://autobuild.buildroot.org/results/89f39dae01c07977062b909079bfcd45a68060d3

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

package/htpdate: fix build with static openssl

Depending on its build configuration, openssl might not be compiled with
dynamic zlib support, and programs that link against openssl need to link
against zlib as well.

This can be generally and easily handled using the pkg-config files
installed by openssl.

Fixes: http://autobuild.buildroot.net/results/ae9/ae946ca72238840b3eaa5fe823e8d620618f7462/
Signed-off-by: Titouan Christophe <titouan.christophe@railnova.eu>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/ghostscript: security bump to version 9.53.0

- Use tar.gz as SHA512SUMS does not contain the hash for tar.xz
- Fix CVE-2020-15900: A memory corruption issue was found in Artifex
  Ghostscript 9.50 and 9.52. Use of a non-standard PostScript operator
  can allow overriding of file access controls. The 'rsearch'
  calculation for the 'post' size resulted in a size that was too large,
  and could underflow to max uint32_t.

https://www.ghostscript.com/doc/9.53.0/News.htm

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

package/re2: needs threads

re2 unconditionally uses -pthread and add it to re2.pc

Moreover, it fails to buid without threads on:

In file included from /home/peko/autobuild/instance-0/output-1/per-package/re2/host/opt/ext-toolchain/arm-buildroot-linux-uclibcgnueabihf/include/c++/8.3.0/arm-buildroot-linux-uclibcgnueabihf/bits/os_defines.h:39,
                 from /home/peko/autobuild/instance-0/output-1/per-package/re2/host/opt/ext-toolchain/arm-buildroot-linux-uclibcgnueabihf/include/c++/8.3.0/arm-buildroot-linux-uclibcgnueabihf/bits/c++config.h:508,
                 from /home/peko/autobuild/instance-0/output-1/per-package/re2/host/opt/ext-toolchain/arm-buildroot-linux-uclibcgnueabihf/include/c++/8.3.0/bits/stl_algobase.h:59,
                 from /home/peko/autobuild/instance-0/output-1/per-package/re2/host/opt/ext-toolchain/arm-buildroot-linux-uclibcgnueabihf/include/c++/8.3.0/memory:62,
                 from /home/peko/autobuild/instance-0/output-1/build/re2-2020-08-01/re2/filtered_re2.h:24,
                 from /home/peko/autobuild/instance-0/output-1/build/re2-2020-08-01/re2/filtered_re2.cc:5:
/home/peko/autobuild/instance-0/output-1/per-package/re2/host/arm-buildroot-linux-uclibcgnueabihf/sysroot/usr/include/features.h:218:5: warning: #warning requested reentrant code, but thread support was disabled [-Wcpp]
 #   warning requested reentrant code, but thread support was disabled
     ^~~~~~~
In file included from /home/peko/autobuild/instance-0/output-1/build/re2-2020-08-01/re2/prog.h:22,
                 from /home/peko/autobuild/instance-0/output-1/build/re2-2020-08-01/re2/bitstate.cc:28:
/home/peko/autobuild/instance-0/output-1/build/re2-2020-08-01/re2/re2.h:768:16: error: 'once_flag' in namespace 'std' does not name a type
   mutable std::once_flag rprog_once_;
                ^~~~~~~~~
/home/peko/autobuild/instance-0/output-1/build/re2-2020-08-01/re2/re2.h:768:11: note: 'std::once_flag' is defined in header '<mutex>'; did you forget to '#include <mutex>'?
/home/peko/autobuild/instance-0/output-1/build/re2-2020-08-01/re2/re2.h:218:1:
+#include <mutex>

Fixes:
 - http://autobuild.buildroot.org/results/7d7c6dcac3cb8ea6deb753178e85eb1c5c74c8e3

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Reviewed-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/memtester: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/e1c276603475a952847fd53724e56101cdccbbcb

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

package/motion: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/99d99dad219776e88aaedcaaf38eb4530d0870ed

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

package/libssh: security bump to version 0.9.5

- Drop patches (already in version)
- Fix CVE-2020-16135: libssh 0.9.4 has a NULL pointer dereference in
  tftpserver.c if ssh_buffer_new returns NULL.

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

package/libnetconf2: drop patch

With updating libssh to version 0.9.5 this patch is no longer needed.

http://lists.busybox.net/pipermail/buildroot/2020-September/291986.html

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

package/python: do not install MS Windows installers on the target

Buildroot generates Linux systems, so there is no need to have
MS Windows executables in there.

This reduces the target filesystem size by about 600kB:
$ du -hcs usr/lib/python2.7/distutils/command/wininst-*.exe
60K usr/lib/python2.7/distutils/command/wininst-6.0.exe
64K usr/lib/python2.7/distutils/command/wininst-7.1.exe
60K usr/lib/python2.7/distutils/command/wininst-8.0.exe
192K    usr/lib/python2.7/distutils/command/wininst-9.0.exe
220K    usr/lib/python2.7/distutils/command/wininst-9.0-amd64.exe
596K    total

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

package/python3: do not install MS Windows installers on the target

Buildroot generates Linux systems, so there is no need to have
MS Windows executables in there.

This reduces the target filesystem size by about 2MB:
$ du -hcs usr/lib/python3.8/distutils/command/wininst-*.exe
60K usr/lib/python3.8/distutils/command/wininst-6.0.exe
64K usr/lib/python3.8/distutils/command/wininst-7.1.exe
60K usr/lib/python3.8/distutils/command/wininst-8.0.exe
192K    usr/lib/python3.8/distutils/command/wininst-9.0.exe
220K    usr/lib/python3.8/distutils/command/wininst-9.0-amd64.exe
188K    usr/lib/python3.8/distutils/command/wininst-10.0.exe
220K    usr/lib/python3.8/distutils/command/wininst-10.0-amd64.exe
448K    usr/lib/python3.8/distutils/command/wininst-14.0.exe
576K    usr/lib/python3.8/distutils/command/wininst-14.0-amd64.exe
2,0M    total

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

package/librsvg: bump to version 2.48.8

- Add a dependency to host-rustc
- libcroco is not a dependency since
  https://gitlab.gnome.org/GNOME/librsvg/-/commit/03ce9bd7875ef2a91979bc4f7d6fa5188cfd785c
- gtk3 is not a dependency since
  https://gitlab.gnome.org/GNOME/librsvg/-/commit/522aeee0ca9b4ff3940bea2d13ebb77ec6c064a6
- cairo script is mandatory since switch to rust

This bump is needed to remove libcroco which has been archived and won't
get any security updates anymore:
https://gitlab.gnome.org/Archive/libcroco/-/issues/8

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

package/cairo: enable script for host

script support on host-cairo is needed for host-librsvg

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

package/cairo: enable gobject for host

gobject is needed for host-librsvg

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

package/go: bump to 1.15.2

go1.15.2 (released 2020/09/09) includes fixes to the compiler, runtime,
documentation, the go command, and the net/mail, os, sync, and testing packages.

https://golang.org/doc/devel/release.html#go1.15

Signed-off-by: Christian Stewart <christian@paral.in>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/babeltrace2: disable -Werror

Add an upstream patch that disable -Werror to avoid the following build
failure:

lttng-live.c: In function 'lttng_live_msg_iter_next':
lttng-live.c:1226:15: error: 'stream_iter' may be used uninitialized in this function [-Werror=maybe-uninitialized]
     int ret = common_muxing_compare_messages(
               ^
lttng-live.c:1198:38: note: 'stream_iter' was declared here
   struct lttng_live_stream_iterator *stream_iter;
                                      ^
lttng-live.c: At top level:
cc1: error: unrecognized command line option '-Wno-cast-function-type' [-Werror]
cc1: all warnings being treated as errors

Fixes:
 - http://autobuild.buildroot.org/results/ff41326f913e71f982b3d5cbaf054f8be47913aa

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

package/xerces: fix typo, which re-enables package

Commit 4e60247be67a8e7f9937a8bf2b123ddb2be9dad2 ("package/xerces: fix
coding style in Config.in") introduced a silly typo in the xerces
package depends on... which effectively made it impossible to enable
this package as the "USE_WCHAR" option does not exist.

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

package/haveged: change systemd service file to run earlier

Drop default dependencies, haveged needs nothing but local sockets and
/dev/random.

The service file now mostly matches the upstream Fedora file, except a
lot of isolation options have been dropped. The benefit for a
completely controlled system is small, and those option would pull in
dependencies, delaying entropy being filled up.

Signed-off-by: Norbert Lange <nolange79@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

support/gnuconfig: bump version

Fixes issues with the triple on IBM s390x and Z machines.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
[yann.morin.1998@free.fr: update both, using the update script]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/scripts/pycompile: add --verbose option

Add a new option that prints the (runtime) path of compiled .py files
when VERBOSE=1 is set.

Signed-off-by: Robin Jarry <robin.jarry@6wind.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/scripts/pycompile: fix .pyc original source file paths

When generating a .pyc file, the original .py source file path is
encoded in it. It is used for various purposes: traceback generation,
.pyc file comparison with its .py source, and code inspection.

By default, the source path used when invoking compileall is encoded in
the .pyc file. Since we use paths relative to TARGET_DIR, we end up with
paths that are only valid when relative to '/' encoded in the installed
.pyc files on the target.

This breaks code inspection at runtime since the original source path
will be invalid unless the code is executed from '/'.

Unfortunately, compileall cannot be forced to use the proper path. It
was not written with cross-compilation usage in mind.

Rework the script to call py_compile.compile() directly with pertinent
options:

- The script now has a new --strip-root argument. This argument is
  optional but will always be specified when compiling py files in
  buildroot.
- All other (non-optional) arguments are folders in which all
  "importable" .py files will be compiled to .pyc.
- Using --strip-root=$(TARGET_DIR), the future runtime path of each .py
  file is computed and encoded into the compiled .pyc.

No need to change directory before running the script anymore.

The trickery used to handle error reporting was only applicable with
compileall. Since we implement our own "compileall", error reporting
becomes trivial.

Previously, we had a --force option to tell compileall.compiledir() to
forcibly recompile files if they had changed. Now, we would have to
handle it ourselves. It turns out to not be easy and would need us to
delve into the format of bytecompiled files to extract metadata and
compare it with the expected values, that being even dependent on the
python version being used (fortunately, only two for us: python 2.7 and
the latext 3.x).

Still, this is deemed too complex, and byte-compiling is pretty fast, so
much so that it should be eclipsed by the build duration anyway.

So we just drop support for --force, and instead we always byte-compile.

Signed-off-by: Julien Floret <julien.floret@6wind.com>
Signed-off-by: Robin Jarry <robin.jarry@6wind.com>
[yann.morin.1998@free.fr:
  - always byte-compile
  - drop --force
  - expand commit log to state so and explain why
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

configs/kontron_smarc_sal28: new defconfig

Now that upstream supports this board add buildroot support for it,
too.

This supports:
 - latest upstream linux kernel. Since the support for this board and
   its SoC is still in active development, we want to follow the
   upstream kernel releases for now.
 - building an image which can be installed to eMMC, SD card or USB
   thumb drive.
 - upstream u-boot patches are pending, therefore no support for
   building a bootloader for now.

Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Heiko Thiery <heiko.thiery@gmail.com>
Tested-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/babeltrace2: new package

Babeltrace 2 is a trace manipulation toolkit.

The Babeltrace 2 project offers a library with a C API, Python 3
bindings, and a command-line tool which makes it very easy for mere
mortals to view, convert, transform, and analyze traces.

See <https://babeltrace.org/> for more details.

Babeltrace 2 is a major update of Babeltrace 1 (Buildroot package
`lttng-babeltrace`). Both projects are coinstallable. Except for the
command-line tool (named `babeltrace2`), the Babeltrace 2 project is not
backward compatible with Babeltrace 1.

I'm naming this package `babeltrace2` instead of `lttng-babeltrace2`
because, although it can read LTTng traces, the two projects are
independent. All major distributions use `babeltrace2` as the
Babeltrace 2 package's name.

I'm keeping the `lttng-babeltrace` package because, as of this date, we
still add bug and security fixes from time to time, therefore the
project is not in EOL stage. Some external, custom packages could still
depend on the Babeltrace 1 library, for example.

As with `lttng-babeltrace`, you can build and install the host version
of Babeltrace 2 for the workflow where you trace the target, download
the resulting trace (or receive it during the tracing process), and then
read and analyze it with Babeltrace 2.

If you enable the `elfutils` package (`BR2_PACKAGE_ELFUTILS`), then
support for Babeltrace 2's debugging information filter component
class is enabled.

Tested with glibc, uClibc-ng, and musl.

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

package/weston-imx: remove deprecated package

- the package couldn't be selected with vivante backend anyway since
  the split
- the package couldn't be updated to newer versions as it would imply
  forking libdrm and wayland-protocols for it to build
- drawback of not using NXP weston is the lack of G2D support, let's
  have it as a known issue for Buildroot but use standard weston
  package

Signed-off-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/weston: add imx-gpu-viv as possible egl provider

As the imx-gpu-viv package provides both EGL and gbm libraries.

Signed-off-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/harfbuzz: change to meson build

- add defaults for all meson options (and order according to
  meson_options.txt file)
- drop pthread linking flags previously needed for the test
  programs (maybe not needed by meson, tests disabled by option)

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

package/harfbuzz: bump version to 2.7.2

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

package/harfbuzz: bump to version 2.6.8

- update home page URL (old one is redirected to github)
- change download to new location (github)
- update license hash because of copyright year update:

diff COPING:
-Copyright © 2010,2011,2012,2013,2014,2015,2016,2017,2018,2019  Google, Inc.
-Copyright © 2019  Facebook, Inc.
+Copyright © 2010,2011,2012,2013,2014,2015,2016,2017,2018,2019,2020  Google, Inc.
+Copyright © 2018,2019,2020  Ebrahim Byagowi
+Copyright © 2019,2020  Facebook, Inc.

note: newer versions (2.7.x) require a migration to meson
Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/nginx-upload: bump to latest version

Update nginx-upload to a newer Git commit which adds upload support over
HTTP2 as well as other fixes.

Remove a patch which is no longer required in the current version.

Signed-off-by: Robert Hancock <hancock@sedsystems.ca>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/graphicsmagick: add support for webp image format

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

board/olpc/xo-1.75: enable drivers that were added in 5.8

This notably makes sound work.

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/olpc_xo175_defconfig: bump kernel version to 5.8.2

Update kernel to version 5.8.2.

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
[Peter: explicitly select 5.8 headers]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/iptables: fix static build

Static build of iptables 1.8.5 fails on:

/srv/storage/autobuild/run/instance-3/output-1/host/opt/ext-toolchain/bin/../lib/gcc/arm-buildroot-linux-uclibcgnueabi/8.3.0/../../../../arm-buildroot-linux-uclibcgnueabi/bin/ld: ../libxtables/.libs/libxtables.a(libxtables_la-xtables.o): in function `dlreg_free':
/srv/storage/autobuild/run/instance-3/output-1/build/iptables-1.8.5/libxtables/xtables.c:235: undefined reference to `dlclose'

Issue has been added with
https://git.netfilter.org/iptables/commit/?id=7db4333dc0b6cd8e943fab78d6dab40ff9f4512e

Fixes:
 - http://autobuild.buildroot.org/results/00f551be87507757a78e332f2859b1e7b3190363

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

package/subversion: drop first patch

First patch is not needed since version 1.8.0 and
https://github.com/apache/subversion/commit/f071ec0c26cdf47e89fa90b31d2233ee1a2b00c2

Indeed, as spotted by upstream when sending them this patch, the
original expressions will not mangle '-mfloat-gprs=double' because the
patterns contain a trailing space.

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

boot/afboot-stm32: fix build with gcc 10

This commit adds a patch that passes -fno-builtin to fix the build
with gcc 10. The patch has been submitted upstream.

There are no autobuilder failures, since it's a bootloader package,
and therefore not tested by the autobuilders. However, a build failure
was detected when building toolchains and a minimal ARMv7-M system at
https://gitlab.com/bootlin/toolchains-builder/-/jobs/729359529.

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

package/opentyrian: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/241b594bab63b2eeceb749e01510da550ed0ffd1

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

package/trinity: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/bac2cbc791e2ab1d0995c8417bd3d19cbc884e47

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

package/musepack: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/9971da8c59484a6a26eb79d6e157461a47855dbf

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

package/libcap: fix static linking issues

Since the bump of libcap to 2.42, host-libcap unconditionally tries to
build a shared library, which fails on build machines where the static
version of the C library is not available.

This issue was reported upstream, who fixed it by two different
commits, which are backported as patches 0001 and 0002. They require
passing a DYNAMIC= value, which should be "yes" to enable dynamic
linking, or empty when not using dynamic linking.

However, other upstream changes broke our logic to support static-only
linking for the target. So we introduce a 0003 patch which extends how
the DYNAMIC flag is used to disable the build of the shared library in
the libcap/ folder. This allows to greatly simplify libcap.mk.

Note that the refactoring is mixed with the fix: the two are hardly
splitable. We need to be able to pass the same options at all steps, and
especially the staging step, otherwise some code gets compiled with the
host compiler, installed in staging, and thus fails the architecture
check later on.

Fixes:

  host-libcap build failure on system without a static libc
  http://autobuild.buildroot.net/results/4b14458014e420ffe088f118e7d0261e67f2d551/

  libcap build failure on static only systems
  http://autobuild.buildroot.net/results/8961759067c4639ae697b6ee5db606f098b7c7e8/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[yann.morin.1998@free.fr:
  - also pass DYNAMIC=yes at host-install step
  - extend commit log to explain why we refactor and fix together
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/wlroots: always disable xcb-errors

Stop trying to enable the xcb-errors Meson option and always disable it:
there is no xcb-util-errors package in Buildroot. The build recipe was
wrongly assuming that the XCB errors library would be installed as part
of the xcb-util package, which is not the case.

Fixes the following autobuilder failure:

  http://autobuild.buildroot.net/results/1891aa624b29e4bbcbbe6bc752a1cc90c316024e/

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libxml-parser-perl: use the compiler as "LD"

Since commit efa95b19aeab6bdcd7d5a620e199c2c32cf2516b
("package/libxml-parser-perl: make host build use correct compiler"),
we pass $(HOST_CONFIGURE_OPTS) when building host-libxml-parser-perl,
in order to use the correct host compiler.

However, this means that LD="$(HOSTLD)" is passed. However, the
host-libxml-parser-perl passes compiler arguments to LD, so it really
assumes that LD is gcc, not ld. For example, it tries to pass
-mtune=generic.

So, let's tell host-libxml-parser-perl that LD is "$(HOSTCC)".

Fixes:

  http://autobuild.buildroot.net/results/2ed2e5ccefe9047c597f84d5880de2e8de2bdd94/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Tested-by: Joseph Kogut <joseph.kogut@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/iwd: bump to 1.9

Also drop mainlined [1] patch.

[1] https://git.kernel.org/pub/scm/network/wireless/iwd.git/commit/?id=6bf5c3ddbd1e8f50e8c06bf8ff3cda79ee436de4

Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Reviewed-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/ell: bump to 0.33

Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Reviewed-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/neardal: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/7efb100c899b67ffd570f73c202442f95ca5397e

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/boost: bump to version 1.74.0 to fix C++20

GCC 10 with -std=c++20 breaks boost library due to allocator changes.
https://github.com/boostorg/bimap/issues/23

We need boost 1.74.0 to fix build with gcc-10 -std=c++20. Note that none
of the Buildroot packages that use boost set -std=c++20, so no in-tree
failures.

Signed-off-by: Maxim Kochetkov <fido_max@inbox.ru>
Reviewed-by: Michael Nosthoff <buildroot@heine.tech>
Tested-by: Michael Nosthoff <buildroot@heine.tech>  for arm
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/libcap: disable GOLANG

Recently, the go infrastructure has changed, and libcap has been
upgraded to 2.42. libcap introduced golang in 2.28, see:

https://git.kernel.org/pub/scm/linux/kernel/git/morgan/libcap.git/commit/Make.Rules?id=0615d996379dceedefcd65a114f93fefd81c208f

When you compile host-go and then compile host-libcap, GOLANG will be
automatically set to yes. Because libcap.mk lacks golang support,
compilation will fail.

Signed-off-by: Tian Yuanhao <tianyuanhao@aliyun.com>
Tested-by: Joseph Kogut <joseph.kogut@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/scripts/pycompile: sort imports

Signed-off-by: Robin Jarry <robin.jarry@6wind.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/scripts/pycompile: add main entry point

Only run code when the script is executed directly (not imported).
Factorize command description by using the script's __doc__ variable.
Fix typo in --force help message.

Signed-off-by: Robin Jarry <robin.jarry@6wind.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/linux-headers: the latest headers are 5.8

In commit c2009e9f757fe759c74406f1025967b7b8e77f0a
("package/linux-headers: license files hashes only valid for latest
version"), we introduced BR2_KERNEL_HEADERS_LATEST, which should only
be set for the most recent kernel headers versions.

Indeed, the COPYING file of Linux has changed before/after Linux 5.6,
causing its hash file to be different. Since linux-headers uses
linux/linux.hash as the hash file, and this hash file contains the
COPYING hash of Linux >= 5.6, we cannot use that hash for Linux
versions older than 5.6.

When newer versions of the headers than 5.4 were added, this
BR2_KERNEL_HEADERS_LATEST was not moved as it should have been. We fix
this, which fixes a legal-info failure happening when Linux kernel
headers 5.4 are used:

>>> linux-headers 5.4.63 Patching
>>> linux-headers 5.4.63 Collecting legal info
ERROR: COPYING has wrong sha256 hash:
ERROR: expected: fb5a425bd3b3cd6071a3a9aff9909a859e7c1158d54d32e07658398cd67eb6a0
ERROR: got     : ee5808b032a67f587d3541099d46de34f5bec8cd5976114ba07f1299ee6001ff
ERROR: Incomplete download, or man-in-the-middle (MITM) attack

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

gitlab-ci: only check defconfigs for known branches

Currently, the check of defconfigs is run for all branches, even those
that are pushed only to run runtime tests. This is very inconvenient.

In fact, we only want to check the defconfigs on standard branches, that
is master, next, and the maintenance branches.

This will also decrease drastically the number gitlab-ci minutes used
when one pushes their repo to gitlab.com, where the number of CI minutes
are now going to be pretty severely restricted.

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

gitlab-ci: move before-script before script

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

gitlab-ci: handle the basic tests in script

Note that those tests were so far ignored only when requesting a single
defconfig build, or a single runtime test build; everything else
was trigerring thoses tests.

However, it feels more natural that they are also ignored when all
defconfigs build. or all runtime tests, are explictly requested.

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

gitlab-ci: handle single runtime test in script

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

gitlab-ci: handle single defconfig in script

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

gitlab-ci: handle the run-time tests conditions in script

Note that we do not propagate the existing comment, because it is
partially wrong; instead we just keep the per-condition comments.

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

gitlab-ci: defconfig_check is included in defconfig build

When we build the defconfigs, we already check they are correct, so
there is no need to run the correctness check explicitly.

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

gitlab-ci: handle the defconfig build conditions in script

Note that we do not propagate the existing comment, because it is
partially wrong; instead we just keep the per-condition comments.

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

gitlab-ci: share the image version with the child

Currently, the image name and version are duplicated in the main
pipeline and the generated, child pipeline.

This is a condition for a future gaffe, so let's use the image from the
main pipeline when generating the child one.

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

gitlab-ci: introduce main() in generating script

This script is currently very crude, but we're going to extend it, at
which point it will be nicer to have functions, local variables, et al.

Introduce a main() in preparation of those future evolutions.

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

package/zeromq: security bump to version 4.3.3

Fixes the following security issues:

- CVE-2020-15166: Denial-of-Service on CURVE/ZAP-protected servers by
  unauthenticated clients.
  If a raw TCP socket is opened and connected to an endpoint that is fully
  configured with CURVE/ZAP, legitimate clients will not be able to exchange
  any message.  Handshakes complete successfully, and messages are delivered
  to the library, but the server application never receives them.  For more
  information see the security advisory:
  https://github.com/zeromq/libzmq/security/advisories/GHSA-25wp-cf8g-938m

- Stack overflow on server running PUB/XPUB socket (CURVE disabled).
  The PUB/XPUB subscription store (mtrie) is traversed using recursive
  function calls.  In the remove (unsubscription) case, the recursive calls
  are NOT tail calls, so even with optimizations the stack grows linearly
  with the length of a subscription topic.  Topics are under the control of
  remote clients - they can send a subscription to arbitrary length topics.
  An attacker can thus cause a server to create an mtrie sufficiently large
  such that, when unsubscribing, traversal will cause a stack overflow.  For
  more information see the security advisory:
  https://github.com/zeromq/libzmq/security/advisories/GHSA-qq65-x72m-9wr8

- Memory leak in PUB server induced by malicious client(s) without CURVE/ZAP.
  Messages with metadata are never processed by PUB sockets, but the
  metadata is kept referenced in the PUB object and never freed.  For more
  information see the security advisory:
  https://github.com/zeromq/libzmq/security/advisories/GHSA-4p5v-h92w-6wxw

- Memory leak in client induced by malicious server(s) without CURVE/ZAP.
  When a pipe processes a delimiter and is already not in active state but
  still has an unfinished message, the message is leaked.
  For more information see the security advisory:
  https://github.com/zeromq/libzmq/security/advisories/GHSA-wfr2-29gj-5w87

- Heap overflow when receiving malformed ZMTP v1 packets (CURVE disabled).
  By crafting a packet which is not valid ZMTP v2/v3, and which has two
  messages larger than 8192 bytes, the decoder can be tricked into changing
  the recorded size of the 8192 bytes static buffer, which then gets
  overflown by the next message.  The content that gets written in the
  overflown memory is entirely decided by the sender.
  For more information see the security advisory:
  https://github.com/zeromq/libzmq/security/advisories/GHSA-fc3w-qxf5-7hp6

Drop now upstreamed patches, autoreconf and reformat hash file with 2 space
delimiters.

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

docs/manual: Add section about contributing to maintenance branches

Signed-off-by: Brandon Maier <brandon.maier@rockwellcollins.com>
[yann.morin.1998@free.fr:
  - s/release branch/maintenance branch/
  - extend the master-then-backport section
  - slight eye-candy on the rest
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

configs/stm32mp157: remove extra metadata_csum mkfs option

Since commit [1] in U-Boot upstrea, there is no necessity to pass extra
metadata_csum option due to changed env location in U-Boot so we can
drop it completely.

[1] 76db1681da52342ca9f4fb7e6787bd83cc82f429:
stm32mp1: use a specific SD/eMMC partition for U-Boot enviromnent

Signed-off-by: Bartosz Bilas <b.bilas@grinn-global.com>
[yann.morin.1998@free.fr: 76db1681d referecnes a U-Boot commit]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/kbd: needs gcc 4.9

kbd uses _Generic since version 2.2.90 and
https://github.com/legionus/kbd/commit/da5feb8fd90c3a0df9d505d4604c629d051d3d7d

However, _Generic is not available until gcc 4.9:
https://gcc.gnu.org/wiki/C11Status

As a result, build with gcc 4.8 fails on:

setleds.c:352:3: warning: implicit declaration of function '_Generic' [-Wimplicit-function-declaration]
   ndefflags = BITMASK_SET(BITMASK_UNSET(odefflags, ndef), nval);
   ^
setleds.c:22:2: error: expected expression before 'unsigned'
  unsigned char: (unsigned char)(~(x)) \
  ^

Fixes:
 - http://autobuild.buildroot.org/results/b74ecdda44543da1d47fa2c027fb046a3ca1e2d1

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[yann.morin.1998@free.fr: propagate the MMU dependency to the comment]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/wlroots: drop libcap dependency

libcap has been drroped since version 0.11.0 and
https://github.com/swaywm/wlroots/commit/906c0766df2fec4bd32c316fd1b0d46fded5fc84

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Reviewed-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/iptables: bump to version 1.8.5

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/python-ipdb: new package

IPython-enabled pdb.

This package can be very helpful when someone wants to debug
a Python application on the board itself.

Signed-off-by: Asaf Kahlon <asafka7@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/strace: disable mpers support

On aarch64 With the config option "--enable-mpers=check" the configure.ac
script searchs for a 32bit compiler. When a matching compiler is found
in the PATH some compatiblity checks are done. This can fail when the
available kernel headers on host and buildroot target does not match.

Since buildroot does not support 32bit binaries when building for 64bit
architecture (no -m32 option) we can disable this option unconditionally.

When disabling unconditionally also the configuration for toolchain using
MUSL can be removed.

Cc: Baruch Siach <baruch@tkos.co.il>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Brandon Maier <brandon.maier@rockwellcollins.com>
Signed-off-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/uhubctl: bump to version 2.2.0

Force GIT_VERSION to avoid git version logic that takes the Buildroot
git commit id.

Update LICENSE hash; copyright year update.

Format hashes file with two space delimiters.

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