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>

package/wireguard-linux-compat: bump version to 0.0.20200215

Fixes a regression introduced in 0.0.20200214.  For details, see the
announcement:
https://lists.zx2c4.com/pipermail/wireguard/2020-February/005014.html

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

package/libgpg-error: bump to version 1.37

- Remove 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/python-cython: bump to version 0.29.15

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/python-simplejson: bump to version 3.17.0

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/python-pyyaml: bump to version 5.3

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/python-pyopenssl: bump to version 19.1.0

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/gensio: bump to version 1.5.1

- Update indentation of hash file (2 spaces)
- This will fix a build failure without threads thanks to
  https://github.com/cminyard/gensio/commit/8918de5b30f90b826c48064e9ee92304b63ffe85
  and associated upstream patch

Fixes:
 - http://autobuild.buildroot.org/results/e94d0e0b46afc1223a74bcc471909f4adef0d6f3

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

package/libtorrent-rasterbar: bump to version 1.2.4

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/python-six: bump to version 1.14.0

License hash change is due to date update.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/python-cryptography: bump to version 2.8

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/wpewebkit: security bump to version 2.26.4

Fixes the following security issues:

- CVE-2020-3862: Impact: A malicious website may be able to cause a denial
  of service.  Description: A denial of service issue was addressed with
  improved memory handling.

- CVE-2020-3864: Impact: A DOM object context may not have had a unique
  security origin.  Description: A logic issue was addressed with improved
  validation.

- CVE-2020-3865: Impact: A top-level DOM object context may have incorrectly
  been considered secure.  Description: A logic issue was addressed with
  improved validation.

- CVE-2020-3867: Impact: Processing maliciously crafted web content may lead
  to universal cross site scripting.  Description: A logic issue was
  addressed with improved state management.

- CVE-2020-3868: Impact: Processing maliciously crafted web content may lead
  to arbitrary code execution.  Description: Multiple memory corruption
  issues were addressed with improved memory handling.

For more details, see the advisory:
https://wpewebkit.org/security/WSA-2020-0002.html

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>

package/wpewebkit: needs >= GCC 7

CMakeLists.txt contains a toolchain check:

if (${CMAKE_CXX_COMPILER_ID} STREQUAL "GNU")
    if (${CMAKE_CXX_COMPILER_VERSION} VERSION_LESS "7.3.0")
        message(FATAL_ERROR "GCC 7.3 or newer is required to build WebKit. Use a newer GCC version or Clang.")
    endif ()
endif ()

So bump the toolchain dependency to >= GCC 7.  The check is really about >=
7.3.0, but we do not have such detailed version checks.  Given that GCC
7.3.0 was released in January 2018 (and 7.1.0 in May 2017), most external
GCC 7.x toolchains probably use >= 7.3.0.

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

package/webkitgtk: security bump to version 2.26.4

Fixes the following security issues:

- CVE-2020-3862: Impact: A malicious website may be able to cause a denial
  of service.  Description: A denial of service issue was addressed with
  improved memory handling.

- CVE-2020-3864: Impact: A DOM object context may not have had a unique
  security origin.  Description: A logic issue was addressed with improved
  validation.

- CVE-2020-3865: Impact: A top-level DOM object context may have incorrectly
  been considered secure.  Description: A logic issue was addressed with
  improved validation.

- CVE-2020-3867: Impact: Processing maliciously crafted web content may lead
  to universal cross site scripting.  Description: A logic issue was
  addressed with improved state management.

- CVE-2020-3868: Impact: Processing maliciously crafted web content may lead
  to arbitrary code execution.  Description: Multiple memory corruption
  issues were addressed with improved memory handling.

For more details, see the advisory:
https://webkitgtk.org/security/WSA-2020-0002.html

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>

package/webkitgtk: needs >= GCC 7

CMakeLists.txt contains a toolchain check:

if (${CMAKE_CXX_COMPILER_ID} STREQUAL "GNU")
    if (${CMAKE_CXX_COMPILER_VERSION} VERSION_LESS "7.3.0")
        message(FATAL_ERROR "GCC 7.3 or newer is required to build WebKit. Use a newer GCC version or Clang.")
    endif ()
endif ()

So bump the toolchain dependency to >= GCC 7.  The check is really about >=
7.3.0, but we do not have such detailed version checks.  Given that GCC
7.3.0 was released in January 2018 (and 7.1.0 in May 2017), most external
GCC 7.x toolchains probably use >= 7.3.0.

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

package/libcurl: rename curl binary config symbol

Package optional or choice config symbols are usually prefixed with the
package config symbol name. Rename BR2_PACKAGE_CURL to
BR2_PACKAGE_LIBCURL_CURL to conform.

Update references to the old name.

Cc: Matt Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Reviewed-by: Matt Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/mfgtools: fix build issue related to __time64_t

The tool fails to build on recent distros due to conflicting declaration
of __time64_t. Adding a check around the declaration to avoid
redefinition.

Patch not submitted upstream as the tool is not supported by NXP
anymore[1].

Fixes:
http://autobuild.buildroot.net/results/ca4498ad21a96ba2a38ca2467dadffdbb516355b/

[1] https://github.com/NXPmicro/mfgtools/pull/104

Signed-off-by: Gary Bisson <bisson.gary@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

docs/manual: describe the new <pkg>_IGNORE_CVES variable

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

support/scripts/pkg-stats: add support for CVE reporting

This commit extends the pkg-stats script to grab information about the
CVEs affecting the Buildroot packages.

To do so, it downloads the NVD database from
https://nvd.nist.gov/vuln/data-feeds in JSON format, and processes the
JSON file to determine which of our packages is affected by which
CVE. The information is then displayed in both the HTML output and the
JSON output of pkg-stats.

To use this feature, you have to pass the new --nvd-path option,
pointing to a writable directory where pkg-stats will store the NVD
database. If the local database is less than 24 hours old, it will not
re-download it. If it is more than 24 hours old, it will re-download
only the files that have really been updated by upstream NVD.

Packages can use the newly introduced <pkg>_IGNORE_CVES variable to
tell pkg-stats that some CVEs should be ignored: it can be because a
patch we have is fixing the CVE, or because the CVE doesn't apply in
our case.

>From an implementation point of view:

 - A new class CVE implement most of the required functionalities:
   - Downloading the yearly NVD files
   - Reading and extracting relevant data from these files
   - Matching Packages against a CVE

 - The statistics are extended with the total number of CVEs, and the
   total number of packages that have at least one CVE pending.

 - The HTML output is extended with these new details. There are no
   changes to the code generating the JSON output because the existing
   code is smart enough to automatically expose the new information.

This development is a collective effort with Titouan Christophe
<titouan.christophe@railnova.eu> and Thomas De Schampheleire
<thomas.de_schampheleire@nokia.com>.

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

package/{mesa3d, mesa3d-headers}: bump version to 19.3.4

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

package/rocksdb: add gflags optional dependency

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

package/mono: fix build with powerpc

Fixes:
 - http://autobuild.buildroot.org/results/fff0dd08f71facbe367d982d19158ee084ae8047

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

package/wireguard-linux-compat: bump version to 0.0.20200214

Includes misc fixes. For details, see the announcement:
https://lists.zx2c4.com/pipermail/wireguard/2020-February/005013.html

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

package/postgresql: security bump to version 12.2

Fixes the following security issues:

- CVE-2020-1720: ALTER ... DEPENDS ON EXTENSION is missing authorization checks
  https://www.postgresql.org/about/news/2011/

Update the license hash for a change in copyright years:
-Portions Copyright (c) 1996-2019, PostgreSQL Global Development Group
+Portions Copyright (c) 1996-2020, PostgreSQL Global Development Group

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

package/screen: bump version to 4.8.0

Fixes a memory corruption issue in OSC 49 handling.  Notice that this is
only enabled if screen is built with --enable-rxvt_osc, which isn't the case
in Buildroot. From the release notes:

As last fix, fixes potential memory overwrite of quite big size (~768
bytes), and even though I'm not sure about potential exploitability of
that issue, I highly recommend everyone to upgrade as soon as possible.
This issue is present at least since v.4.2.0 (haven't checked earlier).

https://lists.gnu.org/archive/html/screen-devel/2020-02/msg00007.html

Upstream changed the gnu.org URLs to use HTTPS, so adjust
0005-rename-sched_h.patch to match.

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

DEVELOPERS: add Romain Naour for toolchain topic

The first time I worked on the Buildroot's toolchain infra
was to add support for the Sourcery Codebench Standard
(licenced) edition toolchain (from Mentor Graphics) for
x86 target [1]. The series was rejected though.

But the knowledge gained from this work served to refactor
the toolchain-external infra in Buildroot [2].

Nowadays, I'm using toolchains-builder project to do
some toolchain build testing to keep GNU tools up to date
in Buildroot.

[1] http://lists.busybox.net/pipermail/buildroot/2014-November/112036.html
[2] http://lists.busybox.net/pipermail/buildroot/2016-October/175433.html
[3] https://gitlab.com/kubu93/toolchains-builder/

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

DEVELOPERS: add Romain Naour for Qemu defconfigs

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

DEVELOPERS: add Romain Naour for test_glxinfo test

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

support/testing/glxinfo: explicitely enable GLX

Since [1], the GLX support is enabled by BR2_PACKAGE_MESA3D_OPENGL_GLX
symbol.

Since [2], only one swrast provider can be built.
Keep BR2_PACKAGE_MESA3D_DRI_DRIVER_SWRAST.

Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/400391349

[1] 5cb821d5635626b7327d5d704555c412e5ed5a1f
[2] 09a0a285076f544de335efc74c8904e464576575

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

package/ncdu: bump to version 1.14.2

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

package/libmicrohttpd: bump to version 0.9.70

Bugfix release. For details, see the release notes:
https://lists.gnu.org/archive/html/libmicrohttpd/2020-02/msg00006.html

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

package/libhttpparser: bump to version 2.9.3

Also dropped patch that was pushed upstream

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

package/go: bump version to 1.13.8

Includes fixes to the runtime, the crypto/x509, and net/http
packages.

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

package/dovecot: security bump to version 2.3.9.3

Fixes the following security issues:

- CVE-2020-7046: Truncated UTF-8 can be used to DoS submission-login and
  lmtp processes
  lib-smtp doesn't handle truncated command parameters properly, resulting
  in infinite loop taking 100% CPU for the process.  This happens for LMTP
  (where it doesn't matter so much) and also for submission-login where
  unauthenticated users can trigger it.

- CVE-2020-7957: Specially crafted mail can crash snippet generation
  Snippet generation crashes if:
  - message is large enough that message-parser returns multiple body
    blocks
  - The first block(s) don't contain the full snippet (e.g.  full of
    whitespace)
  - input ends with '>'

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

package/parted: disable on uclibc

Like postgreSQL (and imagemagick), parted does not build against uClibc
with locales enabled, due to an uClibc bug, see
http://lists.uclibc.org/pipermail/uclibc/2014-April/048326.html:

In file included from atari.c:42:
atari.c: In function 'atr_part_correct':
atari.c:221:9: error: dereferencing pointer to incomplete type 'struct __uclibc_locale_struct'
  return isalnum_l(part->id[0], atr_c_locale)
         ^~~~~~~~~

So disable parted on uclibc

Fixes:
 - http://autobuild.buildroot.org/results/992518d340a9f32a0721d6e66936850c4c3ef2e4

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

package/udisks: add locale dependency

Commit b5f0c6efb24826641719c493382211e5d768417b forgot to propagate new
locale dependency from parted to udisks

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

package/python-pyparted: add locale dependency

Commit b5f0c6efb24826641719c493382211e5d768417b forgot to propagate new
locale dependency from parted to python-pyparted

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

package/libsigrok: drop remnants of autoreconf

libsigrok has not needed autoreconf since b428801934 (package/libsigrok:
bump version to 0.4.0), 4 years ago now.

As such, we no longer need the autoreconf options, nor the dependency on
the autoconf archive.

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Cc: Bartosz Golaszewski <brgl@bgdev.pl>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Tested-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/sdl2: fix build without threads

- Drop first patch (not needed since bump to version 2.0.10 and
  https://github.com/spurious/SDL-mirror/commit/2601ef1f2d7f998c1d276d1b06cac4ed7feba2e1)
- Add a new patch to fix an outstanding build failure

Fixes:
 - http://autobuild.buildroot.org/results/7f7712c5bd47de4a3fcec1e0d0526fd5a3ecd532

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

package/eudev: Fix monitor starting for kernels w/o CONFIG_SHMEM

When the kernel has CONFIG_SHMEM disabled, /dev is a ramfs (instead of a
tmpfs) and the name_to_handle_at system call is not supported. This
causes eudev's monitor application to exit on startup.

Upstream eudev has added this fix which is not yet part of a release.

Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/kodi-screensaver-asteroids: bump version to 2.3.2

Switched _LICENSE_FILES to debian/copyright.

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

package/kodi-peripheral-steamcontroller: bump version

Switched _LICENSE_FILES to debian/copyright.

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

package/kodi-peripheral-joystick: bump version to 1.4.9-Leia

Switched _LICENSE_FILES to debian/copyright.

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

package/kodi-inputstream-rtmp: bump version to 2.0.8-Leia

Switched _LICENSE_FILES to debian/copyright.

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

package/rocksdb: switch to generic-package

Switch from cmake-package to generic-package to allow rocksdb to run the
./build_tools/build_detect_platform script and detect compiler options
such as C++17 support for -faligned-new

First patch needs to be updated and second patch can be dropped

Fixes:
 - http://autobuild.buildroot.org/results/22c9909c0d20e3871775f3874f7723910d7e5a41
 - http://autobuild.buildroot.org/results/ab7b2bc9e9653a7093d8b27d4445c28993572ca4

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

package/sqlcipher: enable back libressl

libressl support has been fixed since version 3.4.2 and
https://github.com/sqlcipher/sqlcipher/commit/ce489ebb4788207f27b1641f8d2bfe6b65462260

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

package/sqlcipher: security bump to version 4.3.0

>From https://www.zetetic.net/blog/2019/08/14/defcon-sqlite-attacks:

"We strongly recommend that all applications upgrade to SQLCipher 4.2.0
to take advantage of the latest security updates, especially if an
application interacts with non-encrypted databases using SQLCipher."

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

package/bootstrap: security bump to version 4.3.1

- Fix CVE-2018-14042: In Bootstrap before 4.1.2, XSS is possible in the
  data-container property of tooltip.
- Fix an XSS vulnerability (CVE-2019-8331) in our tooltip and popover
  plugins by implementing a new HTML sanitizer
- 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/squashfs: update homepage link

Development moved to github.com.

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

package/doxygen: needs host gcc >= 4.9

host-doxygen use std::make_unique which is a C++14 feature and so not
available with host gcc 4.8 so add a Config.in.host for doxygen and add
host gcc 4.9 dependency to host-doxygen and sigrok C++ option

Fixes:
 - http://autobuild.buildroot.org/results/3ac78c5d4728287bafdfeb3a54f50eb193934b63

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

package/modem-manager: bump to version 1.12.6

https://lists.freedesktop.org/archives/modemmanager-devel/2020-February/007713.html

Signed-off-by: Aleksander Morgado <aleksander@aleksander.es>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/libxmlpp: bump to version 3.2.0

- Update site in Config.in, see
  https://gitlab.gnome.org/GNOME/libxmlplusplus/commit/604ae3c286a20683f9688f8da5221b3bf89886d0
- 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/qpdf: bump to version 9.1.1

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/mongoose: security bump to version 6.17

- Fix CVE-2019-19307: An integer overflow in parse_mqtt in mongoose.c in
  Cesanta Mongoose 6.16 allows an attacker to achieve remote DoS
  (infinite loop), or possibly cause an out-of-bounds write, by sending
  a crafted MQTT protocol packet.
- 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/libxml-parser-perl: bump to version 2.46

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

package/iozone: bump to version 3_489

Also install "fileop", another file system benchmarking tool
provided by the iozone package.

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

package/sdl2_mixer: bump to version 2.0.4

Also enabled support for Opus music playback using opusfile library

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

package/rng-tools: bump version to 6.9

Fixes high rngd startup latency along with other minor bugs:
https://github.com/nhorman/rng-tools/releases/tag/v6.9

Signed-off-by: Wesley Chow <wes.chow@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/ogre: link with libatomic when needed

On some architectures, atomic binutils are provided by the libatomic
library from gcc. Linking with libatomic is therefore necessary,
otherwise the build fails with:

/home/test/autobuild/run/instance-1/output-1/host/lib/gcc/xtensa-buildroot-linux-uclibc/8.3.0/../../../../xtensa-buildroot-linux-uclibc/bin/ld: ../../lib/libOgreMain.so.1.12.0: undefined reference to `__atomic_fetch_add_8'

This is often for example the case on sparc v8 32 bits.

Fixes:
 - http://autobuild.buildroot.org/results/3a09e2d1d26b19243244eb7f9235c85488a788d2

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

package/mesa3d: fix circular dependency in Kconfig

When the r300 driver was introduced in c5ae77c97 (package/mesa3d: add
support for gallium r300 driver), a last-minute fix was introduced by
Yann, to properly propagate the dependency of a selected symbol.

However, this ended up causing a spurious circular dependency that does
not really exists, but that Kconfig is not smart enough to detect is in
fact OK.

Fixing this is pretty non-obvious, but we have an easy way out: the
dependency is about libdrm's radeon driver requirement for a toolchain
that has the sync4 family of primitives, which is always a given for an
x86 toolchain. As the radeon r300 driver is x86-only, this dependency is
forcefully fulfilled.

So, we drop the propagated dependency, and replace it by a fat comment.

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Cc: Romain Naour <romain.naour@gmail.com>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Cc: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

boot/arm-trusted-firmware: add missing qstrip

Add missing qstrip wrapping to the new
BR2_TARGET_ARM_TRUSTED_FIRMWARE_ADDITIONAL_TARGETS option.

Signed-off-by: Francois Gervais <fgervais@distech-controls.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/wpewebkit: disable JSC JIT for MIPSr6

Forcibly disable the JavaScriptCore JIT compilation support
for MIPSr6 processors, which are unsupported.

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

package/webkitgtk: disable JSC JIT for MIPSr6

Forcibly disable the JavaScriptCore JIT compilation support
for MIPSr6 processors, which are unsupported.

Fixes: http://autobuild.buildroot.net/results/3d21d3c3460cd85a4c828dd197929cdf17aaf4e0
Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/musl: move riscv64 register index constant definitions

The riscv64 build of libsigsegv using musl fails due to a missing
definition for REG_SP. This constant is used to index the __gregs
array in the ucontext_t structure.

This fix moves the musl defintion of REG_SP (and others) from
arch/riscv64/bits/reg.h to arch/riscv64/bits/signal.h
so that it is picked up correctly.

The patch was downloaded from upstream:
https://git.musl-libc.org/cgit/musl/commit/?id=329e79299daaa994b8e75941331a1093051ea5d9

Fixes:
http://autobuild.buildroot.org/results/8fcd5cb912e513ac08a81e3ee726e29ac22212bb/

Signed-off-by: Mark Corbin <mark@dibsco.co.uk>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/exim: fix build error during install step

exim builds some files during the 'make install' step, and these fail with
an error:

  lookups/lf_quote.c:49:3: error: 'for' loop initial declarations are only allowed in C99 mode
     for (int j = 0; j < vlength; j++)
     ^

Fix by passing the -std=c99 here, as it is already passed in the build
step.

Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/exim: fix various build failures

Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/exim: fix target build on some toolchains

Building with the Sourcery CodeBench ARM 2014.05 the build fails with this
error:

  >>> exim_dbmbuild utility built

  .../buildroot/output/host/bin/arm-none-linux-gnueabi-gcc -DEXIM_DUMPDB exim_dbutil.c
  exim_dbutil.c: In function 'main':
  exim_dbutil.c:568:1: error: 'for' loop initial declarations are only allowed in C99 mode
   for (uschar * key = dbfn_scan(dbm, TRUE, &cursor);
   ^
  exim_dbutil.c:568:1: note: use option -std=c99 or -std=gnu99 to compile your code
  exim_dbutil.c:630:2: error: 'for' loop initial declarations are only allowed in C99 mode
    for (int i = 1; i <= wait->count; i++)
    ^
  exim_dbutil.c:642:6: error: 'for' loop initial declarations are only allowed in C99 mode
        for (int j = 0; j < MESSAGE_ID_LENGTH; j++)
        ^

Fix by enforcing C99. This completes commit
2c692e81a844b30b4d3161dfd9897b3265bb9279 ("package/exim: fix host build")
to also fix target builds.

Fixes: http://autobuild.buildroot.net/results/6b7e08090f5f0f2627cc3e89b349c2052b6e3116/
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Config.in.legacy: fix selection of bluez5_utils

Select BR2_PACKAGE_BLUEZ5_UTILS only if all its reverse dependencies
are selected

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

package/gqrx: bump to version 2.12.1

- Drop all patches (already in version)
- Update indentation in hash file (two spaces)
- This bump also fix a build failure with BR2_ENABLE_DEBUG thanks to
  https://github.com/csete/gqrx/commit/d5f8e008253d63b71f5b2cf41651214f9a78e83c

Fixes:
 - http://autobuild.buildroot.org/results/17e564fc6465e6e83742c421f2a48b8a0a4923bc

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

package/php-*: fix autoconf variables

Build of php-* packages are broken since commit
3292f3de499717360cf18d7fecd79ac096216ca8 because
$(HOST_DIR)/bin/auto{conf,header} have been replaced by
$(HOST_DIR)/bin/auto{conf,header} -I "$(ACLOCAL_DIR)" -I "$(ACLOCAL_HOST_DIR)"

So revert this change (which should have been only refactoring) for
those packages

Fixes:
 - http://autobuild.buildroot.org/results/0d5f4a792652e302b8c00c1e0f99966cc930fbac
 - http://autobuild.buildroot.org/results/91a559183db212803a90cb399a1cb505b14818d9

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

package/uclibc: restore __isctype_l definition

Recent is*_l fix broke uclibc build because removed __isctype_l
definition was used in libc/misc/ctype/ctype.c. Restore it.

Fixes: 8723c5e7a6db ("package/uclibc: fix ctype.h is*_l definitions")
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
[yann.morin.1998@free.fr:
  - add new patch, don't fix existing one
  - add URL to upstream ML post
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/mesa3d: add support for gallium r300 driver

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
[yann.morin.1998@free.fr: propagate libdrm's radeon dependencies]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libexif: add post-0.6.21 upstream security fixes

Fixes the following security issues:

- CVE-2016-6328: A vulnerability was found in libexif.  An integer overflow
  when parsing the MNOTE entry data of the input file.  This can cause
  Denial-of-Service (DoS) and Information Disclosure (disclosing some
  critical heap chunk metadata, even other applications' private data).

- CVE-2017-7544: libexif through 0.6.21 is vulnerable to out-of-bounds heap
  read vulnerability in exif_data_save_data_entry function in
  libexif/exif-data.c caused by improper length computation of the allocated
  data of an ExifMnote entry which can cause denial-of-service or possibly
  information disclosure.

- CVE-2018-20030: An error when processing the EXIF_IFD_INTEROPERABILITY and
  EXIF_IFD_EXIF tags within libexif version 0.6.21 can be exploited to
  exhaust available CPU resources.

- CVE-2019-9278: In libexif, there is a possible out of bounds write due to
  an integer overflow.  This could lead to remote escalation of privilege in
  the media content provider with no additional execution privileges needed.
  User interaction is needed for exploitation.

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

package/gnuradio: backport build fixes with less use of Boost

Backport patch from upstream to fix build failures such as:

    In file included from /home/buildroot/autobuild/instance-0/output-1/build/gnuradio-3.8.0.0/gr-digital/lib/glfsr.cc:23:
    /home/buildroot/autobuild/instance-0/output-1/build/gnuradio-3.8.0.0/gr-digital/lib/../include/gnuradio/digital/glfsr.h:42:5: error: 'uint32_t' does not name a type; did you mean 'u_int32_t'?
        uint32_t d_shift_register;
        ^~~~~~~~
        u_int32_t

Since Gnuradio policy is Less boost == better and C++11 is used, use cstdint
instead of boost/cstdint.hpp.

Applied in gnuradio master (475e4a156b516c089175afb998acdc80b740b437)

fix:
 - http://autobuild.buildroot.net/results/14015f499e58fee530877ac052878bbe2f799942/
 - http://autobuild.buildroot.net/results/53239f98dd5e03d4dc1bb4eb91ed765f77dbf0ec/

Signed-off-by: Gwenhael Goavec-Merou <gwenhael.goavec-merou@trabucayre.com>
[yann.morin.1998@free.fr:
  - add upstream reference in the patch itself
  - minor eye-candy in commit log
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/mesa3d: add support for gallium i915 driver

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
[yann.morin.1998@free.fr: also guard comment with x86 dependency]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/mesa3d: fix iris driver depends and selects

iris is inherently an x86-only driver, and it hard codes gcc options
specific to x86m like -msse2, causing build breakage on other
architectures.

iris also does not use kmsro, but the select was accidentally added when
iris was introduced.

Fix both by adding the missing dependency to x86, and by removing the
select to kmsro.

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
[yann.morin.1998@free.fr:
  - ad dependency to x86
  - reword commit log accordingly
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/weston: update patches with upstream backports

These updated patches fix the same issues but are backported from upstream
commits instead of pull requests.

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

package/Config.in: fix make failure by removing gstreamer/Config.in

After commit:
https://git.buildroot.net/buildroot/commit/?id=fb49c7a26182f9d48f8283e7328fddc216962c94
gstreamer entry in package/Config.in was left behind resulting in every
make call to fail. So let's remove orphaned gstreamer entry from
package/Config.in

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/qemu: do not support x86_steamroller or x86_core_avx2

These CPU's cause segfaults with qemu.

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

package/gstreamer/*: remove packages

Gstreamer 0.10 has been deprecated upstream since 2012 and is missing a lot
of features and (security) fixes compared to gstreamer1, so remove it.

All gstreamer-0.10 sub packages depends on gstreamer, so we only need to add
a legacy entry for that.

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

package/nvidia-tegra23-binaries: drop gstreamer 0.10.x support

With the upcoming removal of gstreamer 0.10, the logic for installing
binaries using gstreamer 0.10.x in nvidia-tegra23-binaries must go as well.

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

package/freerdp: drop gstreamer 0.10.x support

With the upcoming removal of gstreamer 0.10, the logic for building freerdp
with support for it must go as well.

As there is now a single option for gstreamer (1.x) support, convert the
gstreamer support choice to a normal option for simplicity.

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

package/opencv3: drop gstreamer 0.10.x support

With the upcoming removal of gstreamer 0.10, the logic for building opencv3
with support for it must go as well.

As there is now a single option for gstreamer (1.x) support, convert the
gstreamer support choice to a normal option for simplicity.

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

package/opencv: drop gstreamer 0.10.x support

With the upcoming removal of gstreamer 0.10, the logic for building opencv
with support for it must go as well.

As there is now a single option for gstreamer (1.x) support, convert the
gstreamer support choice to a normal option for simplicity.

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

package/libplayer: remove package

Libplayer is dead upstream.  The mercurial repo is no longer online, it
hasn't seen any releases since 2010 and the mplayer backend was removed from
Buildroot in 2018.

With the upcoming removal of gstreamer 0.10, there is no longer any backends
available in Buildroot, so remove the package.

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