package/iprutils: fix build with ancient toolchains

alphasort and uint used to be guarded behind various test-macros
definitions, which are not default on ancient toolchains (glibc-2.18
era for example).

Fixes:
 - http://autobuild.buildroot.org/results/41fde4aa06f7c025bb05aa594716643b9010358f

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[yann.morin.1998@free.fr: issue is with glibc, not gcc]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/ibm-sw-tpm2: fix glibc build errors

In order to not have to define all architecture bitness variants in
src/LibSupport.h to conditionally define RADIX_BITS, a patch was created
to depend on LONG_BIT being defined via limits.h

For glibc, LONG_BIT is behind a _XOPEN_SOURCE define guard so patch the
makefile to define this macro as well as _DEFAULT_SOURCE.

Defining _XOPEN_SOURCE requires _DEFAULT_SOURCE to be explicitly defined
for builds to succeed as it is no longer impliclty defined when
_XOPEN_SOURCE is defined.

Fixes:
http://autobuild.buildroot.net/results/bc145dd23c80a0660300766d0c114a00ed2e52b6/
http://autobuild.buildroot.net/results/1d84e70c30d4d0719274aa4f3edc01772beb24a3/

Fixes: 19bd089004 ("package/ibm-sw-tpm2: fix patch to address musl builds")
Signed-off-by: Vincent Fazio <vfazio@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/libunwind: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/dad61a90a27a97f275efd238ee22be704eab0063

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

package/ghostscript: security bump to version 9.52

Fix a bunch of CVEs: CVE-2020-16287, CVE-2020-16288, CVE-2020-16289,
CVE-2020-16290, CVE-2020-16291, CVE-2020-16292, CVE-2020-16293,
CVE-2020-16294, CVE-2020-16295, CVE-2020-16296, CVE-2020-16297,
CVE-2020-16298, CVE-2020-16299, CVE-2020-16300, CVE-2020-16301,
CVE-2020-16302, CVE-2020-16303, CVE-2020-16304, CVE-2020-16305
CVE-2020-16308, CVE-2020-16309, CVE-2020-17538

PKGCONFIG must be passed since version 9.51 and
https://github.com/ArtifexSoftware/ghostpdl/commit/2d84ecc57837785b566ebd9d5909ba9edc9d697f

Also drop patch (already in version) and update indentation in hash file
(two spaces)

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

package/lttng-tools: fix build with musl

Fixes:
 - http://autobuild.buildroot.org/results/0c49b3e90361ced63cad4807dbe1b13c595f5f97

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

package/ifplugd: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/63f7c6fc4a2a30c8eb0298856a2615f816d50299

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

package/iftop: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/55a8581f11e776439782c228441ef1c8c1243386

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

package/collectd: add buddyinfo option

buddyinfo is available since version 5.11.0 and
https://github.com/collectd/collectd/commit/cb19ecf90d89a79836b8b5236d2b66277dd79c5f

It is currently enabled by default

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Reviewed-by: Asaf Kahlon <asafka7@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/collectd: add capabilities option

capabilities plugin is available since version 5.11.0 and
https://github.com/collectd/collectd/commit/b934f1572cdb54cf485b600546a72dbf9929ba2c

It is currently enabled by default depending on jansson and
libmicrohttpd availability

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Reviewed-by: Asaf Kahlon <asafka7@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/collectd: add InfluxDB support

InfluxDB support is available since version 5.11.0 and
https://github.com/collectd/collectd/commit/605fcdcddeefb23b945203079e5c87625ce0e336

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Reviewed-by: Asaf Kahlon <asafka7@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/collectd: add DPDK_TELEMETRY option

DPDK support is available since version 5.11.0 and
https://github.com/collectd/collectd/commit/00cbd4d2c016a676db476ae66bbc58d6bf11a6a0

It depends on jansson and is currently enabled by default raising the
following build failure on musl:

src/dpdk_telemetry.c:43:10: fatal error: sys/unistd.h: No such file or directory
 #include <sys/unistd.h>
          ^~~~~~~~~~~~~~

Fixes:
 - http://autobuild.buildroot.org/results/aafb8c72f147fefc7a988c45e4dc17de48b07a95

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

package/libcamera: disable werror

This will fix the following build failure:

../src/gstreamer/gstlibcameraallocator.cpp: In static member function 'static GQuark FrameWrap::getQuark()':
/home/peko/autobuild/instance-1/output-1/host/x86_64-buildroot-linux-gnu/sysroot/usr/include/glib-2.0/glib/gatomic.h:128:15: error: variable 'gapg_temp_atomic' set but not used [-Werror=unused-but-set-variable]
     gpointer *gapg_temp_atomic = (gpointer *)(atomic);                       \
               ^

Fixes:
 - http://autobuild.buildroot.org/results/c28500d4cc55fbd2bac87f2c11759ddc9163bc91

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

package/pistache: fix detection of atomic

pistache does not mandatory link with libatomic since
https://github.com/oktal/pistache/commit/4ed3f8e148dae203730139ae3c1c4c87bfc86c02
however the logic is broken because the C language is not enabled

Fixes:
 - http://autobuild.buildroot.org/results/17f05238a3479f5b0c9ef8f9bb42210f6fc14fff

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

package/opencv3: fix build with jasper >= 2.0.17

Fixes:
 - http://autobuild.buildroot.org/results/8da00d4b079195f45fe74d879b10db05d74d0559

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

package/opencv: fix build with jasper >= 2.0.17

Fixes:
 - http://autobuild.buildroot.org/results/656e2232a0566ba8f7826a87b1fab9cc2c3d8e46

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

package/fluidsynth: fix build with gcc 4.8

Fixes:
 - http://autobuild.buildroot.org/results/13cbba871db56ef8657a3d13c6ac8e1b4da0d244

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

Revert "package/libnss: enable parallel building"

This reverts commit 31602d8aaabaa9f55048c3db7a1b1b2fb2d27ef2 as it seems
that libnss still have parallel build issues.

Fixes:
 - http://autobuild.buildroot.org/results/521aaf8b41686865ea86688ea6410418b422fd55

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

package/bandwidthd: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/6308c8ee38b6017215038d47c009b238113bd36f

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

package/mpv: atomic is mandatory

__sync builtins have been dropped since version 0.24.0 and
https://github.com/mpv-player/mpv/commit/c3205d294e519f994829d9355eb133670729aab4

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

package/bluez-alsa: update gcc 10 patch

Update patch by adding the missing update on src/ofono-iface.h from
https://github.com/Arkq/bluez-alsa/commit/30a23dd7ce9b00c702fa48545f6b03038f9e17c6

Fixes:
 - http://autobuild.buildroot.org/results/9200bdab254e24ecacd345c00216a3d87c014dfc

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

package/mpv: link with libatomic if needed

Fixes:
 - http://autobuild.buildroot.org/results/12f4580429427b6b546184366f74da16f83d692c

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

package/ibm-sw-tpm2: fix patch to address musl builds

Previously, __WORDSIZE was being used to define RADIX_BITS.

__WORDSIZE is not defined consistently via limits.h across all libc
variants, so use LONG_BIT instead.

Fixes:

  http://autobuild.buildroot.net/results/b5cb347a98dc64f77982633c0fc48d14030aa5b6/

Signed-off-by: Vincent Fazio <vfazio@gmail.com>
Signed-off-by: Vincent Fazio <vfazio@xes-inc.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/keepalived: ipset support only makes sense when iptables is enabled

In keepalived, ipset is only used when iptables support is
enabled. The configure.ac script is quite convoluted, but one can
clearly see:

AS_IF([test .$enable_iptables != .no],
  [
   ... testing for iptables ...
    if test $USE_IPTABLES = Yes; then
      dnl ----[Check for ipset libraries]----
      SAV_CPPFLAGS="$CPPFLAGS"
      CPPFLAGS="$CPPFLAGS $kernelinc"
      if test "${enable_libipset}" != no; then
      ... testing for libipset ...
    fi
  ]

So, the libipset library is only tested and used if iptables is
enabled. This is also visible by the code that shows at the end of the
configure script which features are enabled:

echo "Use iptables             : ${USE_IPTABLES}"
if test .$USE_IPTABLES = .Yes; then
  echo "Use libipset             : ${USE_LIBIPSET}"
fi

Once again, the libipset support only makes sense when iptables is
enabled.

However, the configure.ac script also has some logic to detect if a
--<something>-libipset option is passed while iptables is enabled:

AS_IF([test .$enable_iptables = .no],
    AS_IF([test .$enable_libipset != .], [AC_MSG_ERROR([disable-libipset requires vrrp and iptables])])
  )

The error message is quite misleading because it is in fact displayed
as soon as a --<something>-libipset is passed not just
--disable-libipset.

In the context of Buildroot, we are always passing a
--<something>-libipset, regardless of whether iptables support is
enabled or not, which makes the build error out:

configure: error: disable-libipset requires vrrp and iptables

This commit fixes that by enclosing the libipset logic inside the
iptables condition. When iptables is not available, we pass
--disable-iptables and that's it, nothing else is needed. When
iptables is available, we pass --enable-iptables *and*
--enable-libipset or --disable-libipset depending on the availability
of libipset.

This has been tested successfully with the following combinations:

 - keepalived, without iptables or libipset
 - keepalived, with iptables, but without libipset
 - keepalived, without iptables, but with libipset. In this case
   libipset is obviously not used.
 - keepalived, with iptables and with libipset. Both are used. Note
   that you will not see the keepalived binary linked with libipset.so
   in "readelf -d" because keepalived dlopen()s the libipset.so
   library by default.

Fixes:

  http://autobuild.buildroot.org/results/a1712b2cc3ad878e6876325ec7d4c434d0d9d11b/
  (case with --disable-libipset --disable-iptables)

  http://autobuild.buildroot.net/results/4567e3b0a0510e8a615781178ff5bbbd835a92c3/
  (case with --enable-libipset --disable-iptables)

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

Revert "toolchain/toolchain-buildroot: enable uclibc for riscv64"

This reverts commit bd9810e176273914eca1208bcba23f0de9e446b3. Indeed,
while uClibc-ng has support for RISC-V 64-bit, this support lacks the
__riscv_flush_icache() function call, which is used by some GCC
builtins used for example in libffi.

Due to this missing __riscv_flush_icache(), anything that links
against libffi fails to build:

/home/test/autobuild/run/instance-0/output-1/host/bin/riscv64-buildroot-linux-uclibc-gcc  -o gobject/gobject-query gobject/gobject-query.p/gobject-query.c.o -Wl,--as-needed -Wl,--no-undefined -Wl,-O1 -Wl,--start-group glib/libglib-2.0.so.0.6400.4 gobject/libgobject-2.0.so.0.6400.4 -Wl,--end-group -pthread '-Wl,-rpath,$ORIGIN/../glib:$ORIGIN/' -Wl,-rpath-link,/home/test/autobuild/run/instance-0/output-1/build/libglib2-2.64.4/build/glib -Wl,-rpath-link,/home/test/autobuild/run/instance-0/output-1/build/libglib2-2.64.4/build/gobject
/home/test/autobuild/run/instance-0/output-1/host/lib/gcc/riscv64-buildroot-linux-uclibc/9.3.0/../../../../riscv64-buildroot-linux-uclibc/bin/ld: /home/test/autobuild/run/instance-0/output-1/host/riscv64-buildroot-linux-uclibc/sysroot/usr/lib64/libffi.so.7: undefined reference to `__riscv_flush_icache'
collect2: error: ld returned 1 exit status

Note that this commit means that
support/config-fragments/autobuild/br-riscv64-full-internal.config
will be back to using glibc as the C library, but that is OK, until
uClibc-ng is fixed to implemented __riscv_flush_icache().

This uClibc-ng issue has been reported upstream at
https://mailman.uclibc-ng.org/pipermail/devel/2020-August/002022.html.

Fixes:

  http://autobuild.buildroot.net/results/ec1185ad1fd8863a3990143a0af2ace987761a27/

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

package/pulseview: fix build with recent Qt versions

This commit backports an upstream commit to fix the build of pulseview
with recent Qt versions.

Note that pulseview has a 0.4.2 version (newer than our 0.4.1) but it
doesn't have the fix we need. And 0.4.2 now requires additional Qt
translation tools, so the update to 0.4.2 is anyway not
straightforward.

Fixes:

  http://autobuild.buildroot.net/results/c9e27a91905dfda0ba1ddb12c7910316d3ee82ef/

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

package/bird: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/3e01f490d76be03f21a0680eb162a0c0267fc1b9

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

package/drbd-utils: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/e444f46cd6e6d767592b17b83095737d0508a475

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

package/kmsxx: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/7777cd9496e8f8cdb093c3f82c550eebedda0b5d

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

DEVELOPERS: drop Maxime Ripard from kmsxx maintainers

Maxime Ripard is no longer at Bootlin, his e-mail is bouncing:

<maxime.ripard@bootlin.com>: host spool.mail.gandi.net[217.70.178.1] said: 550
    5.1.1 <maxime.ripard@bootlin.com>: Recipient address rejected: User unknown
    in virtual mailbox table (in reply to RCPT TO command)

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

gitlab-ci: update the image version

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

support/misc/gitlab-ci.yml.in: use python3 for flake8

The pkg-stats script now uses Python3 only constructs (the "async"
keyword) and therefore fails to pass the Python2 flake8 test.

Let's use the Python3 flake8 instead.

Fixes:

  https://gitlab.com/buildroot.org/buildroot/-/jobs/681711009

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: Titouan Christophe <titouan.christophe@railnova.eu>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/docker: drop Python 2.x modules

Since commit 4a40d36f131890941c1e482eb5f28ebcd81cb7a3
("support/testing: switch to Python 3 only") our runtime testing
infrastructure is Python 3.x only.

Therefore, it is no longer needed to have python-nose2 and
python-pexpect in the Docker container used to run our Gitlab CI jobs.

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

configs/qemu_*: bump kernel version to 5.4.58

Bump configs/qemu_* to latest longterm kernel 5.4.42.

All defconfig using a specific kernel version (csky, risc32) are not
taken into account.

Build and runtime tested: https://gitlab.com/jugurthaB/buildroot/-/pipelines/177882458

Signed-off-by: Jugurtha BELKALEM <jugurtha.belkalem@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/redis: bump to version 6.0.6

>From the release notes:

================================================================================
Redis 6.0.6     Released Mon Jul 20 09:31:30 IDT 2020
================================================================================

Upgrade urgency MODERATE: several bugs with moderate impact are fixed here.

The most important issues are listed here:

* Fix crash when enabling CLIENT TRACKING with prefix
* EXEC always fails with EXECABORT and multi-state is cleared
* RESTORE ABSTTL won't store expired keys into the db
* redis-cli better handling of non-pritable key names
* TLS: Ignore client cert when tls-auth-clients off
* Tracking: fix invalidation message on flush
* Notify systemd on Sentinel startup
* Fix crash on a misuse of STRALGO
* Few fixes in module API
* Fix a few rare leaks (STRALGO error misuse, Sentinel)
* Fix a possible invalid access in defrag of scripts (unlikely to cause real harm)

New features:

* LPOS command to search in a list
* Use user+pass for MIGRATE in redis-cli and redis-benchmark in cluster mode
* redis-cli support TLS for --pipe, --rdb and --replica options
* TLS: Session caching configuration support
================================================================================

Also, the reference hashes repo has moved to redis/redis-hashes
probably because the original maintainer has stepped down, see
http://antirez.com/news/133

Finally, the copyright year in the COPYING file has been updated,
and so does its sha256 checksum.

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

package/supertux: fix build with boost 1.73.0

Fixes:
 - http://autobuild.buildroot.org/results/0f02877a665b2281266df4b23c72a7c113906c31

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

package/mpd: add openal optional dependency

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

support/docker: use python3-flake8

support/scripts/pkg-stats now uses some Python 3.x only constructs
("async" and related keywords), so we must use the Python 3.x flake8.

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

utils/scanpypi: use raw strings in re.compile/re.sub

Fixes the following Python 3.x flake8 warning:

W605 invalid escape sequence '\w'

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

utils/getdeveloperlib.py: use raw strings for re.compile/re.match

Fixes the following Python 3.x flake8 warning:

W605 invalid escape sequence '\s'

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

support/testing/tests/core/test_timezone.py: fix indentation

Fixes:

support/testing/tests/core/test_timezone.py:7:9: E117 over-indented

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

support/testing: consistently use raw strings for re.compile

Otherwise Python 3.x flake8 complains with:

W605 invalid escape sequence '\s'

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

utils/{check-package, checkpackagelib}: consistently use raw strings for re.compile

Raw strings need to be used when calling re.compile() otherwise Python
3.x flake8 complains with:

W605 invalid escape sequence '\s'

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

package/18xx-ti-utils: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/df2d04de4b2e5003f3fc56c9e84db5054bd4d499

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

package/cvs: fix mktime related compile failure

Use ac_cv_func_working_mktime=yes to force the use of a provided
mktime implementation instead of compiling the failing own one.

Fixes:

  http://autobuild.buildroot.net/results/5bcd8f4235002da682cc900f866116d2fe87f1c8

  mktime.c: In function 'ydhms_diff':
  mktime.c:106:52: error: size of array 'a' is negative
   #define verify(name, assertion) struct name { char a[(assertion) ? 1 : -1]; }
                                                      ^
  mktime.c:170:3: note: in expansion of macro 'verify'
     verify (long_int_year_and_yday_are_wide_enough,
     ^~~~~~

with the failure/assert comming from the lines:

  verify (long_int_year_and_yday_are_wide_enough,
          INT_MAX <= LONG_MAX / 2 || TIME_T_MAX <= UINT_MAX);

which fails since the y2038 time_t conversion from 32bit to 64bit
(musl libc).

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

package/mpd: add fluidsynth optional dependency

fluidsynth is an optional dependency of mpd since version 0.18.5 and
https://github.com/MusicPlayerDaemon/MPD/commit/727c6226597fb710239b14e7985d97ec6b53f0ee

It is enabled by default

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

package/fluidsynth: needs dynamic library

applications, such as mpd, fail to build statically with fluidsynth
because fluidsynth does not fill Libs.Private in their pkg-config file

Because fluidsynth is a cmake-package with many dependencies, it is not
easy to fix so just add a dynamic library dependency.

Here is an extract of src/CMakeLists.txt that gives the list of
fluidsynth possible dependencies:

target_link_libraries ( libfluidsynth
${GLIB_LIBRARIES}
${GMODULE_LIBRARIES}
${LASH_LIBRARIES}
${JACK_LIBRARIES}
${ALSA_LIBRARIES}
${PULSE_LIBRARIES}
${PORTAUDIO_LIBRARIES}
${LIBSNDFILE_LIBRARIES}
${SDL2_LIBRARIES}
${DBUS_LIBRARIES}
${READLINE_LIBS}
${DART_LIBS}
${COREAUDIO_LIBS}
${COREMIDI_LIBS}
${WINDOWS_LIBS}
${MidiShare_LIBS}
${OpenSLES_LIBS}
${OBOE_LIBS}
${LIBFLUID_LIBS}
${LIBINSTPATCH_LIBRARIES}
)

Fixes:
 - http://autobuild.buildroot.org/results/ec9dd2903359b9bf6b15c8cb69e732f8cb6c4d39

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

package/rtl8821au: add patch fixing sprintf error

Fixes compile errors against certain kernels.

Signed-off-by: Christian Stewart <christian@paral.in>
Reviewed-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/capnproto: add openssl optional dependency

openssl is an optional dependency that is enabled by default since
version 0.7.0 and
https://github.com/capnproto/capnproto/commit/23db5e3fd91104a0b2881d8f8ab3c10bf9dd8e75

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

package/mender-artifact: use github helper for site URL

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

package/mender-artifact: do not download modules during the build process

By default, go will attempt to download needed modules before building, which
is not desirable. This behavior also causes permission issues when cleaning,
as go downloads modules as read-only by default. Because mender-artifact
includes the modules in the vendor directory, mod=vendor prevents the package
from downloading the go modules during the build process and prevents
permission issues when cleaning.

Fixes:
http://autobuild.buildroot.net/results/d5bcaca73ae74fe8b0ebd39b6331564cd639fb66

Signed-off-by: Mirza Krak <mirza.krak@northern.tech>
Signed-off-by: Adam Duskett <Aduskett@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/open-lldp: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/b6aaaa1e72ba9ce0c29730970911431558181c6e

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

package/rauc: fix URL of the signed tarball

Reported-by: Yair Ben Avraham <yairba@protonmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-decorator: drop host variant

Drop host variant of python-decorator as it is not used by any package
since commit d4a0dc5d571cac4f12399d78617db4ff8e55ae12

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

package/qt5webengine: examples need qt5svg

Fixes:

  https://gitlab.com/buildroot.org/buildroot/-/jobs/674934350

  make[3]: Entering directory '.../build/qt5webengine-5.15.0/examples/webenginewidgets'
  Project ERROR: Unknown module(s) in QT: svg
  Makefile:47: recipe for target 'sub-pdfviewer-make_first' failed

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

package/{gnuradio|opencv3|python-matplotlib|qt5|qwt|wireshark}: drop redundant BR2_PACKAGE_QT5BASE select

Drop redundant BR2_PACKAGE_QT5BASE select as BR2_PACKAGE_QT5 already
implies select BR2_PACKAGE_QT5BASE (see package/qt5/Config.in).

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

package/wpa_supplicant: disable TEAP for internal TLS implementation

EAP-TEAP support in hostapd/wpa_supplicant fails to build with internal
TLS implementation. This patch disables TEAP support in wpa_supplicant
when internal TLS implementation is selected. Similar fix for hostapd
package has already been merged to Buildroot: see commit 47d14e3b1c20
("package/hostapd: disable TEAP for internal TLS implementation").

TEAP is still an experimental feature that is not recommmended for
production use. Currently it should not be used for anything else
than experimentation and interoperability testing. Those who needs
experimenting with TEAP are encouraged to enable openssl in their
buildroot configuration.

Fixes:
http://autobuild.buildroot.net/results/e83613c06041a60f89da787f4ebf876245713cd2/

Signed-off-by: Sergey Matyukevich <geomatsi@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python-gunicorn: select python-setuptools

Fixes:
 - https://bugs.buildroot.org/show_bug.cgi?id=13111

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

package/domoticz: fix build with boost 1.73.0

Fixes:
 - http://autobuild.buildroot.org/results/3607112343af24df6ca7bafc660cb5c4927c9008

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

package/ghostscript: 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. This was fixed in commit
5d499272b95a6b890a1397e11d20937de000d31b.

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

package/fluidsynth: bump to version 2.1.4

Changes since v2.1.3:
https://github.com/FluidSynth/fluidsynth/releases/tag/v2.1.4
- fix an uninitialized memory access possibly triggering an FPE trap
- fix several regressions introduced in 2.1.3:
  - fluid_synth_start() failed for certain presets
  - fix a NULL dereference in jack driver
  - fix a stack-based overflow when creating the synth

Signed-off-by: Julien Olivain <juju@cotds.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/openal: bump to version 1.20.1

This will fix the build with gcc 10 due to the library codebase being
switched to C++11.

It should be noted that bumping the C++ requirement from C++11 to C++14
is not needed for this bump as this change is not yet in this version:
https://github.com/kcat/openal-soft/commit/f6a0b004e0424e1f12a226b3b3c7193063eb02e3

Also update indentation in hash file (two spaces)

Fixes:
 - http://autobuild.buildroot.org/results/4546ce4f176164462b0e73c387bc45c9771f98f7

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

package/live555: license is now LGPL-3.0+ and not LGPL-2.1+

The live555 source code includes both a COPYING file (with the GPL-3.0
license text) and a COPYING.LESSER file (with the LGPL-3.0 license
text). However, all source files indicate a LGPL-3.0 license, and none
of them indicate a GPL-3.0 license. In addition,
http://live555.com/liveMedia/faq.html#copyright-and-license says the
source code is under the LGPL.

So, we:

- Bump LGPL License to 3.0+
- Add a comment about the GPL-3.0 license

Fixes:

- https://bugs.busybox.net/show_bug.cgi?id=13156

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

package/python-spidev: switch to setuptools

spidev uses setuptools since version 3.5 and
https://github.com/doceme/py-spidev/commit/3bb2f3b4061f62f283ec2a525e2de60d7f572b41

Fixes:
 - http://autobuild.buildroot.org/results/619b9e9844e33fa8e111d6db09eee8450fedb660

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

package/cpio: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/22fefd9774cbd6648d67f29826f47f1978e9c069

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

package/bellagio: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/0c02f845eabfbc276ee5989520a3c77fc47289ab

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

package/rtl8188eu: bump version to 0924dc8f

- fixes compile against linux-5.4.x

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

package/ima-evm-utils: bump version to 1.3.1

Dropped patches from this release.

List of commits between 1.3 and 1.3.1:

97b912a7277747db1aca74c7d922e37d41b4f9a8 Release version 1.3.1
62534f212798c540c46e9f362e1d84845175d42e Rename "--validate" to "--ignore-violations"
5b58f475708a6a0c2e70a5d98bbe3f530217d1e1 Drop the ima_measurement "--verify" option
8e2738dd44c0aad2e11f995b802a4d6dcde6b505 extend ima_measurement --pcrs option to support per-bank pcr files
79ab82f55f14839d6b9ba6af78af3367370146bb Rename "Changelog" to "NEWS"
7f9a59c6c613ed8da2947b5e0078a8b75c4e7197 Fix missing {u,g}id_t typedef on musl
1f4e423e7c1d8d5ae85e12368065dec2dc4e87fb pcr_tss: Fix compilation for old compilers

Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/iputils: drop wrong linux-headers dependency

Commit 9ffcd9279e5505535a73cf1200b3a86fd425cf32 wrongly added a
linux-headers dependency when switching to meson.

Remove it as headers are always provided by the toolchain.

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

package/python3-pyelftools: use PYTHON3_PYELFTOOLS_VERSION

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

package/python3-mako: use PYTHON3_MAKO_VERSION

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

package/aircrack-ng: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/98ab591a88e06a985b0a7aac5abbf665289134b4

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

package/atest: fix build with gcc 10

Fixes:
 - http://autobuild.buildroot.org/results/887c466b3703449239eedaf86f3f4dd2a2dc8afe

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

package/gdk-pixbuf: security bump to version 2.36.12

- Fix CVE-2017-6312: Integer overflow in io-ico.c in gdk-pixbuf allows
  context-dependent attackers to cause a denial of service (segmentation
  fault and application crash) via a crafted image entry offset in an
  ICO file, which triggers an out-of-bounds read, related to compiler
  optimizations.
- Fix CVE-2017-6313: Integer underflow in the load_resources function in
  io-icns.c in gdk-pixbuf allows context-dependent attackers to cause a
  denial of service (out-of-bounds read and program crash) via a crafted
  image entry size in an ICO file.
- Fix CVE-2017-6314: The make_available_at_least function in io-tiff.c
  in gdk-pixbuf allows context-dependent attackers to cause a denial of
  service (infinite loop) via a large TIFF file.

Also update indentation in hash file (two spaces)

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

package/fping: bump to version 5.0

This bump will fix a build failure with gcc 10 thanks to
https://github.com/schweikert/fping/commit/0ed73181eb54e3f2dc744271c0329620972a85ab

Also update indentation in hash file (two spaces)

Fixes:
 - http://autobuild.buildroot.org/results/f5c8e546200594e2367444e93e31aa615bd73796

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

configs/arcturus_ucls1012a: bump kernel to 4.14.140 and u-boot version to 2019.10

Fixes:

  https://gitlab.com/buildroot.org/buildroot/-/jobs/674933582

Signed-off-by: Michael Durrant <mdurrant@arcturusnetworks.com>
Signed-off-by: Oleksandr G Zhadan <Oleks@ArcturusNetworks.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

configs/ci20: use upstream Linux and U-Boot

The vendor kernel we are currently selecting no longer builds,
and fails with a ton of:

                 from include/linux/list.h:8,
                 from include/linux/module.h:9,
                 from arch/mips/jz4740/prom.c:16:
include/linux/log2.h:22:1: error: ignoring attribute 'noreturn' because it conflicts with attribute 'const' [-Werror=a
ttributes]
   22 | int ____ilog2_NaN(void);
      | ^~~

We can't afford to fix that, so let's just move to upstream
kernel and bootloader. It doesn't make much sense to keep
using an unsupported kernel and bootloader at this point.

This means we will be missing some of the features supported
by the vendor (such as HDMI support), but it is what it is.

Linux v5.7 and v5.4 have been tested to boot fine, the latter
is picked for the defconfig as it is an LTS version.

Fixes:

  https://gitlab.com/buildroot.org/buildroot/-/jobs/674933782

Reported-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/librtlsdr: bump to more recent upstream master version

Since 948666dfde1e0a5d149442f80facc162a83016ad, librtlsdr in Buildroot is no longer
built from an official release, but from a commit on the master branch. However, the
commit that was referenced has a broken pkgconfig file templating, such that
other packages using `pkgconfig --libs librtlsdr` as part of their build process
(such as dump1090) could not be built anymore:

Before 948666dfde1e0a5d149442f80facc162a83016ad:
$ cat staging/usr/lib/pkgconfig/librtlsdr.pc
prefix=/usr
exec_prefix=${prefix}
libdir=${exec_prefix}/lib
includedir=${prefix}/include

Name: RTL-SDR Library
Description: C Utility Library
Version: 0.6.0
Cflags: -I${includedir}/
Libs: -L${libdir} -lrtlsdr -lusb-1.0
Libs.private:

On 948666dfde1e0a5d149442f80facc162a83016ad:
$ cat staging/usr/lib/pkgconfig/librtlsdr.pc
prefix=
exec_prefix=
libdir=
includedir=

Name: RTL-SDR Library
Description: C Utility Library
Version: 7082
Cflags: -I${includedir}/
Libs: -L${libdir} -lrtlsdr
Libs.private:  -lusb-1.0

In the meantime, upstream released a bugfix for that ([1]), so we bump to that
commit as well, and update the only patch for shared libs accordingly, because
upstream also added a new tool called `rtl_biast` in the meantime.

Finaly, we update the hash file to the two-spaces convention.

Fixes: http://autobuild.buildroot.net/results/b4c/b4cdcb59cc61c51c024197a64865ad4b60023d0c/
[1]: https://github.com/steve-m/librtlsdr/commit/ed0317e6a58c098874ac58b769cf2e609c18d9a5

Signed-off-by: Titouan Christophe <titouan.christophe@railnova.eu>
Tested-by: Heiko Thiery <heiko.thiery@gmail.com>
Tested-by: Gwenhael Goavec-Merou <gwenhael.goavec-merou@trabucayre.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

DEVELOPERS: add Gwenhael Goavec-Merou for librtlsdr

Even though librtlsdr was initially introduced by Jason Pruitt in
2014, and Jason is still listed in the DEVELOPERS file for this
package, in recent times it's mainly Gwenhael who has been taking of
this package. Let's reflect that in the DEVELOPERS file so that
Gwenhael gets notified when there are librtlsdr issues.

Cc: Jason Pruitt <jrspruitt@gmail.com>
Cc: Gwenhael Goavec-Merou <gwenhael.goavec-merou@trabucayre.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/librtlsdr: fix version fetching when code is extracted from tarball

librtlsdr currently fails to build on the autobuilders, as it fails
for out of tree builds. Indeed, there is some CMake logic in librtlsdr
that determines the version using Git. This works fine when librtlsdr
is fetched from Git of course. But in the context of Buildroot,
librtlsdr is extracted from a tarball.

For an in-tree build, the "git describe" invocation goes all the way
up to the Buildroot .git/ metadata, and uses that as the librtlsdr
version (it's of course wrong, but the build works). In an out-of-tree
build, there is no parent directory with .git/ metadata, so Git fails,
the VERSION variable is empty and later CMake aborts the build because
of that.

We fix that by adjusting the version retrieving logic to only use Git
if a .git/ metadata folder is found at the root of the librtlsdr
source tree. The patch has been submitted upstream.

Fixes:

  http://autobuild.buildroot.net/results/ea52be1da8ed03272db06679d5a0a441ffe6ea0c/

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

package/htop: add patch for fixing build with gcc 10

Signed-off-by: Maxim Kochetkov <fido_max@inbox.ru>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/x11r7/xserver_xorg-server: add security fix for CVE-2020-14347

Release notes:
https://lists.x.org/archives/xorg-announce/2020-July/003051.html

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

package/python3-decorator: use PYTHON3_DECORATOR_VERSION

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

package/libfuse3: bump to version 3.9.4

Drop patch (already in version). The commit list between 3.9.2 and
3.9.4 is just made of fixes:

db9b4fa148e6c22c0d4b4c567fa65d1cd5368152 Released 3.9.4
50cbca799c1f9b010fabaa0dd4a387f29d140873 fuse_send_data_iov(): correctly calculate total buffer size.
7b3e3899157566875280a8b860eb5ad5c73eadc1 Define fuse_session_loop_mt as a macro on uclibc and MacOS (#532)
c5e8684b5a2f3400af6d7a3edcaeb3ce8ffc51b5 Fixed typo in command to compile program (#536)
e8a9e84672dcaa892d4708c163f768dc177b6d4c Doc fixes (#537)
d1deae6968c49d83334e874c33abfe15824c4548 Fix FreeBSD CI (#539)
48450411647ca0818821af7b05b819ceff92ae7c Fix: crash on failure to set locale (#529)
9e1c2a4959c16c0b50090dd822389ad9acb08111 fuse_lowlevel: Move assert for se before dereferencing it with se->debug (#530)
7471156354002c6547aa6c3a4f39a3262f435ba4 Fixed minor print alignment issue in iconv_help(), replacing tab with space (#519)
9fa4dc1661f085d4e89a54d75acc3347d52f33fa Fix the typo "filed" -> "field" in fuse manpage (#524)
717c8b8b3ed815f14e5607a995d0113446e3fb0b README: Correct the directory name from 'examples' to 'example' (#526)
032db1ab298d62c4d0c5be1f9fb2df299aec2346 docs: Replace `mesonconf` with `meson configure` (#528)
06342ca60ed822b856990915f127d8beddc0d1f6 libfuse: Assign NULL to "old" to avoid free it twice (#522)
5021d6a0a100d4987be126e87b7ee5fbfc17bbdc Typo fixed. (#520)

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

package/mtd: crypto support needs headers >= 4.12

crypto support includes <linux/hash_info.h> since version 2.1.2 and
https://github.com/sigma-star/mtd-utils/commit/a739b59efe7996e3bdcbe8b17743dc05ac7c110a

However, linux/hash_info.h is only exported since kernel 4.12:
https://github.com/torvalds/linux/commit/fcc8487d477a3452a1d0ccbdd4c5e0e1e3cb8bed

Fixes:
 - http://autobuild.buildroot.org/results/dbaeed9e9e2af191166808d2d6d6d3c2e5051c2c

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

support/scripts/pkg-stats: show progress of upstream URL and latest version

This commit slightly improves the output of pkg-stats by showing the
progress of the upstream URL checks and latest version retrieval, on a
package basis:

Checking URL status
[0001/0062] curlpp
[0002/0062] cmocka
[0003/0062] snappy
[0004/0062] nload
[...]
[0060/0062] librtas
[0061/0062] libsilk
[0062/0062] jhead
Getting latest versions ...
[0001/0064] libglob
[0002/0064] perl-http-daemon
[0003/0064] shadowsocks-libev
[...]
[0061/0064] lua-flu
[0062/0064] python-aiohttp-security
[0063/0064] ljlinenoise
[0064/0064] matchbox-lib

Note that the above sample was run on 64 packages. Only 62 packages
appear for the URL status check, because packages that do not have any
URL in their Config.in file, or don't have any Config.in file at all,
are not checked and therefore not accounted.

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

support/scripts/pkg-stats: use aiohttp for upstream URL checking

This commit reworks the code that checks if the upstream URL of each
package (specified by its Config.in file) using the aiohttp
module. This makes the implementation much more elegant, and avoids
the problematic multiprocessing Pool which is causing issues in some
situations.

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

support/scripts/pkg-stats: use aiohttp for latest version retrieval

This commit reworks the code that retrieves the latest upstream
version of each package from release-monitoring.org using the aiohttp
module. This makes the implementation much more elegant, and avoids
the problematic multiprocessing Pool which is causing issues in some
situations.

Since we're now using some async functionality, the script is Python
3.x only, so the shebang is changed to make this clear.

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

package/libabseil-cpp: make check-package happy

The "depends on" must be after the "default" properties.

Fixes:

  https://gitlab.com/buildroot.org/buildroot/-/jobs/677751258

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

support/testing/tests/boot/test_atf: DTC needed for ATF Vexpress test

Since the bump of ATF to 2.2 for the ATF Vexpress test case in commit
fc3d6a3ed0f5f1ceda7ab64b8b2bee88abd29373
("support/testing/tests/boot/test_atf: update U-Boot/ATF use in
TestATFVexpress"), DTC is now needed otherwise the build fails with:

make[2]: dtc: Command not found
Makefile:873: recipe for target 'build/juno/release/fdts/juno_tb_fw_config.dtb' failed

Fixes:

  https://gitlab.com/buildroot.org/buildroot/-/jobs/674934470

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

package/libabseil-cpp: add BR2_PACKAGE_LIBABSEIL_CPP_ARCH_SUPPORTS

The libabseil-cpp package fails to build on a number of CPU
architectures in our autobuilders.

On most CPU architectures, the first issue looked like this:

libabseil-cpp-20200225/absl/base/internal/direct_mmap.h: In function 'void* absl::lts_2020_02_25::base_internal::DirectMmap(void*, size_t, int, int, int, off64_t)':
libabseil-cpp-20200225/absl/base/internal/direct_mmap.h:121:39: error: static assertion failed: Platform is not 64-bit
  121 |   static_assert(sizeof(unsigned long) == 8, "Platform is not 64-bit");
      |                 ~~~~~~~~~~~~~~~~~~~~~~^~~~
libabseil-cpp-20200225/absl/base/internal/direct_mmap.h:123:15: error: 'SYS_mmap' was not declared in this scope; did you mean 'SYS_mmap2'?
  123 |       syscall(SYS_mmap, start, length, prot, flags, fd, offset));
      |               ^~~~~~~~
      |               SYS_mmap2

Indeed, on 32-bit architectures, libabseil-cpp has some special code
to use the mmap2() system call, and it white-lists the supported
architectures. It is therefore trivial to add support for more
architectures.

However, once this is fixed, another issue arises:
absl/debugging/internal/examine_stack.cc uses the ucontext data
structures, which are not provided by uClibc-ng on all CPU
architectures, and even the code of libabseil-cpp does not exist for
all CPU architectures.

So, this commit solves that by simply making libabseil-cpp available
on architectures/C libraries where it is supported: it needs ucontext
support in the toolchain + a CPU architecture where
absl/debugging/internal/examine_stack.cc has the appropriate logic.

This new dependency is propagated to the reverse dependencies of
libabseil-cpp.

With this commit, libabseil-cpp passes a test-pkg -a test (so all
external toolchains used by the autobuilders):

                             andes-nds32 [ 1/45]: SKIPPED
                             arm-aarch64 [ 2/45]: OK
                        br-aarch64-glibc [ 3/45]: OK
                           br-arcle-hs38 [ 4/45]: SKIPPED
                            br-arm-basic [ 5/45]: SKIPPED
                  br-arm-cortex-a9-glibc [ 6/45]: OK
                   br-arm-cortex-a9-musl [ 7/45]: OK
                   br-arm-cortex-m4-full [ 8/45]: SKIPPED
                             br-arm-full [ 9/45]: OK
                    br-arm-full-nothread [10/45]: SKIPPED
                      br-arm-full-static [11/45]: SKIPPED
                   br-i386-pentium4-full [12/45]: OK
                br-i386-pentium-mmx-musl [13/45]: OK
                       br-m68k-5208-full [14/45]: SKIPPED
                      br-m68k-68040-full [15/45]: SKIPPED
                    br-microblazeel-full [16/45]: SKIPPED
                 br-mips32r6-el-hf-glibc [17/45]: OK
                      br-mips64-n64-full [18/45]: OK
                 br-mips64r6-el-hf-glibc [19/45]: OK
                      br-mipsel-o32-full [20/45]: OK
                          br-nios2-glibc [21/45]: SKIPPED
                      br-openrisc-uclibc [22/45]: SKIPPED
               br-powerpc-603e-basic-cpp [23/45]: SKIPPED
             br-powerpc64le-power8-glibc [24/45]: OK
               br-powerpc64-power7-glibc [25/45]: OK
                  br-powerpc-e500mc-full [26/45]: SKIPPED
                              br-riscv32 [27/45]: OK
                              br-riscv64 [28/45]: OK
                         br-riscv64-musl [29/45]: OK
                             br-sh4-full [30/45]: SKIPPED
                        br-sparc64-glibc [31/45]: SKIPPED
                         br-sparc-uclibc [32/45]: SKIPPED
                    br-x86-64-core2-full [33/45]: OK
                          br-x86-64-musl [34/45]: OK
                          br-xtensa-full [35/45]: SKIPPED
                       linaro-aarch64-be [36/45]: OK
                          linaro-aarch64 [37/45]: OK
                              linaro-arm [38/45]: OK
                     sourcery-arm-armv4t [39/45]: OK
                            sourcery-arm [40/45]: OK
                     sourcery-arm-thumb2 [41/45]: OK
                         sourcery-mips64 [42/45]: OK
                           sourcery-mips [43/45]: OK
                          sourcery-nios2 [44/45]: SKIPPED
                         sourcery-x86-64 [45/45]: OK
45 builds, 18 skipped, 0 build failed, 0 legal-info failed

Fixes:

  http://autobuild.buildroot.net/results/ead663b4b67b0b57ed003a46db3182d95cc01bc0/
  (and many similar build failures)

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

package/libabseil-cpp: fix build on RISC-V 32-bits

This commit backports an upstream patch that fixes the build on RISC-V
32-bits.

Fixes:

  http://autobuild.buildroot.net/results/c6c2c99bdba8c840d52fe0464d26aca169ea1d90/

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

package/libfuse3: define fuse_session_loop_mt as a macro on uclibc

This will fix a build failure with fuse-overlayfs

Fixes:
 - http://autobuild.buildroot.org/results/7186515526ee60488dac3bf9c4580ffd13a0ceac

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[yann.morin.1998@free.fr: actually backport upstream commit]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

configs/qemu_arm_vexpress_tz_defconfig: bump ATF version, cleanup config

Bump to a later version of ATF and cleanup the Python
configuration. Previously this configuration had to work around Python
3 issues with OP-TEE. Now this relies on OP-TEE properly building
itself with host-python3.

Signed-off-by: Dick Olsson <hi@senzilla.io>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

boot/optee-os: depend on host-python3

Recent versions of OP-TEE depend on Python 3. Currently, OP-TEE is
building with the Python interpreter provided by the user. This patch
includes an upstream patch that makes the interpreter configurable,
and makes use of this configuration with host-python3.

Signed-off-by: Dick Olsson <hi@senzilla.io>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/python3-pycryptodomex: new package

This package will ensure that pycryptodomex is built for Python 3.
Comments in both python-pycryptodome and python3-pycryptodomex are
added to ensure they stay in sync.

Signed-off-by: Dick Olsson <hi@senzilla.io>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/minizip: fix build with gcc 4.8

Fixes:
 - http://autobuild.buildroot.org/results/b8baff4b86923d29163ce5cc9e391904d81fc90c

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

package/prosody: luabitop is only needed with Lua 5.1

The luabitop package is only available with Lua 5.1. LuaJIT, Lua 5.3
or more recent versions of Lua have this functionality built-in.

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/luabitop: depends on lua 5.1

In preparation for the introduction of Lua 5.4, clarify the dependency
of luabitop: it only makes sense when used with Lua 5.1.

Also update the comment to no longer mention Lua 5.2, since we don't
support Lua 5.2 in Buildroot anymore.

Note that as explained in https://luajit.org/extensions.html, LuaJit
already implements luabitop functions, so luabitop is really for Lua
5.1 only, not for all Lua interpreters that implemented the 5.1 ABI.

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/setools: install for host-python3

setools needs python3 since version 4.2.0 and
https://github.com/SELinuxProject/setools/commit/e292a77c52ed882707c8d165e77914837367313f

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

Fixes:
 - http://autobuild.buildroot.org/results/c2febcea5fbd8a21709721524ae9e9b5fc0896f9

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

package/python3-cython: new package

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

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