yajl: Update to modern standards

samba4: Run init script through shellcheck

youtube-dl: update to version 2019.7.27

ipvsadm: update version to 1.30

Update version to 1.30

Signed-off-by: Florian Eckert <fe@dev.tdt.de>

Signed-off-by: Josef Schlehofer <pepe.schlehofer@gmail.com>

Some minor adjustments.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

Replaced git:// link with https:// which gets through firewalls easier.

Replaced archive with .xz. The one currently in the mirrors has the wrong
hash. .xz is also smaller than .gz.

Eliminated already default CMake option.

Eliminated Build/InstallDev with CMAKE_INSTALL.

Added PKG_BUILD_PARALLEL for faster compilation.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

Some performance improvements by batching file writes.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

The old website still references an old version. Fixes uscan.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

rng-tools: Update to 6.7

While at, update SPDX identifier.

Signed-off-by: Michael Heimpold <mhei@heimpold.de>

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

luaposix: remove eglibc remnants

zstd: Update to 1.4.1

libzmq: Remove uClibc++ support

ykpers: Update to 1.20.0

Signed-off-by: Ken Xu <windedge99@gmail.com>

elektra: Switch to Lua 5.3

After the introduction of Lua 5.3 to OpenWrt, elektra picks it up instead
of 5.1. Since 5.1 is unsupported anyway, switch to 5.3.

Added liblua to PKG_BUILD_DEPENDS as the FindLua script is broken. Having
it here allows compilation to continue.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

intltool: depend on hostpkg first

tar: Add zstd capability

libzmq is used by fbzmq, which absolutely requires libstdcpp since it's a
C++14 project. There's no point in using two libc++.

Cleaned up Makefile for consistency between packages.

Added nanosleep patch for platforms that are missing usleep.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

fastd: fix segmentation fault with musl >1.1.20

When compiled with musl >1.1.20, fastd will crash in case it can't
resolve a peers hostname. This is due to a changed implementation of
freeaddrinfo in musl 1.1.21 onwards.

This segfault is fixed by not calling freeaddrinfo in case the supplied
pointer is null.

Signed-off-by: David Bauer <mail@david-bauer.net>

lmdb: add new package

socat: Fix compilation when usleep is missing

xz: Remove bash dependency

MarkupSafe: Makefile polishing and add src package

zeromq: update to version 4.1.7 (security fix)

expat: update to version 2.2.7 (security fix)

pcre: remove unnecessary dependency on C++ lib

pdns-recursor: bump version to 4.2.0

prometheus: Fix build for mips64/mips64el

kea: remove python3 dependency of kea-admin

python-egenix-mx-base: Makefile polishing

haproxy: Get rid of bashisms from get-latest-patches.sh

Plus some other shellcheck cleanups.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

Bash is not needed here. These scripts are POSIX compliant.

This was originally added as the configure script looks for /bin/bash
locally. Passing gl_cv_posix_shell overrides this.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

Fixed a typo in tcpreplay.

Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>

cifsd

libxslt: patch security issues

Decompression speedup of 7% on average.

Remove upstreamed uClibc-ng patch.

Fixed optimization option to show properly under menuconfig.

Added LTO for smaller size.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

Fixes several issues, including a restart bug and compilation on kernel
4.9.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

Main warning fixed was about echo -e not being supported by POSIX sh.

Added PKG_BUILD_PARALLEL for slightly faster compilation.

Small whitespace cleanups.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

netifyd: Don't build on uClibc-ng

io: Open /dev/mem with O_SYNC for uncached access

Fixes:
CVE-2019-13117
CVE-2019-13118

Signed-off-by: Jan Pavlinec <jan.pavlinec@nic.cz>

Only when /dev/mem is opened with O_SYNC the write and *read* is done
uncached. We saw wrong values read out from the hardware without setting
O_SYNC, the busybox devmem tool showed different values, when O_SYNC is
also set for the io tool, it reads out the same values as devmem.
When looking at the drivers/char/mem.c file in the kernel it is behaving
differently based on the O_DSYNC flag.

Signed-off-by: Hauke Mehrtens <hauke.mehrtens@intel.com>