libkrb5: disable rpath

The default rpath value for krb5 is /usr/lib. On some hosts this might
confuse the build time linker as it looks for target libraries in the
host /usr/lib directory.

Should fix:
http://autobuild.buildroot.net/results/6df/6df84ed325ddd92248741d2f4737423b1b999489/
http://autobuild.buildroot.net/results/049/049f5147288b418402b6e5a4181eff4d426c8d7a/
http://autobuild.buildroot.net/results/432/432d5d0295f77cd34b78e70e8b4b7d77ef19ae3c/

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/xen: force location of init scripts

The Xen configure scripts looks at the build host to decide where to
install the Sys-V startup scripts, and that location differs between
various distros.

Force the location.

Fixes:
    http://autobuild.buildroot.org/results/869/869829ab086e824d164c5c5ec7f087ed83993be6/
    http://autobuild.buildroot.org/results/336/3360e5a9e3d007b4ed77345b5fe93b2dacb6ad49/
    http://autobuild.buildroot.org/results/29e/29e308ce3cc9c83497ba1c1f98fcda3f48fd03c4/

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Matt Weber <matthew.weber@rockwellcollins.com>
Cc: Alistair Francis <alistair@alistair23.me>
Reviewed-by: Alistair Francis <alistair@alistair23.me>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/rauc: add workaround for older kernel headers

Fixes:
    http://autobuild.buildroot.org/results/621/621587906bd2bb27c43b6fcbb051d75f20e32e7c/
    http://autobuild.buildroot.org/results/52c/52ca760f0227075ca988754bb9eb8df0ea3e1563/
    ...

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Andrey Yurovsky <yurovsky@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

libcap: only install shared version

If BR2_SHARED_LIBS is set, only install shared version of library
(continue to build both libraries through all target as there is no
libcap.so target but only a libcap.so.$(VERSION).$(MINOR))

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

DEVELOPERS: Add myself to support/kconfig/

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

package/apr-util: fix ldap detection

configure checks if ldap_set_rebind_proc takes 2 or 3 arguments
http://svn.apache.org/viewvc/apr/apr-util/tags/1.6.1/build/apu-conf.m4?revision=1812528&view=markup#l370

It uses the macro APR_TRY_COMPILE_NO_WARNING which adds -Werror to
treat all warnings as errors when gcc is used:
http://svn.apache.org/viewvc/apr/apr/tags/1.6.3/build/apr_common.m4?revision=1812527&view=markup#l504

In some buildroot configs a compiler warning occurs during this check:
http://autobuild.buildroot.net/results/241/241ed78b93ce86c859e175530fa485711ff61615//apr-util-1.6.1/config.log

/home/rclinux/rc-buildroot-test/scripts/instance-1/output/host/arm-buildroot-linux-uclibcgnueabihf/sysroot/usr/include/features.h:218:5:
 error: #warning requested reentrant code, but thread support was disabled [-Werror=cpp]
cc1: all warnings being treated as errors

Due to this warning 'ac_cv_ldap_set_rebind_proc_style' is set to two
instead of three leading to a build error later on. This patch forces
ac_cv_ldap_set_rebind_proc_style=three to be inline with openldap which
fixes
http://autobuild.buildroot.net/results/241/241ed78b93ce86c859e175530fa485711ff61615/

This solution was inspired by a discussion on the fink mailinglist:
https://sourceforge.net/p/fink/mailman/message/31720482/

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

package/python-cython: fix indentation

Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/67844444
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

log4cplus: add C++11 dependencies

log4cplus uses std::future which is a C++11 feature affected by the gcc
bug 64735

Fixes:
 - http://autobuild.buildroot.net/results/4f4e9ed5dd5bad76eeee5d151e958f22603615b3

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

DEVELOPERS: add myself to some packages

Add myself to lcms2, liblockfile, libnetfilter_conntrack,
libnetfilter_queue and whois

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

linux-headers: bump 4.{9, 14, 16}.x series

Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

linux: bump default to version 4.16.8

Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/qt5cinex: fix download

The Buildroot dl-wrapper tries to download
Qt5_CinematicExperience_%22rpi_%221.0.tgz, which fails, so this commit
fixes the archive name by removing useless double quotes.

Signed-off-by: Julien Boibessot <julien.boibessot@armadeus.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/clang: link libclang against libLLVM.so

libclang.so is statically linking against all LLVM static libraries
instead of linking dynamically against libLLVM.so.

This patch fixes this problem by setting LLVM_LINK_LLVM_DYLIB to ON.

Signed-off-by: Valentin Korenblit <valentin.korenblit@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

python-websockets: backport fix for upstream issue #350

Signed-off-by: Joseph Kogut <joseph.kogut@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

go: security bump to version 1.10.2

This bump contains many bug fixes, as well as the following security
issue, patched in Go 1.10.1:

CVE-2018-7187: The "go get" implementation in Go 1.9.4, when the
-insecure command-line option is used, does not validate the import path
(get/vcs.go only checks for "://" anywhere in the string), which allows
remote attackers to execute arbitrary OS commands via a crafted web
site.

Signed-off-by: Anisse Astier <anisse@astier.eu>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

go: honor building with V=1 for debugging purposes

Signed-off-by: Anisse Astier <anisse@astier.eu>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

go: add self to developers to get cc-ed for build failures

Signed-off-by: Anisse Astier <anisse@astier.eu>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

go: cleanup cross-compilation support

Now that we fixed cross-compilation in the go package, cleanup the build
to remove the workaround added in 60c5c96ae109dcea48e0997b5e6e6645ac9b86
"package/go: Build host tools with host CC". We only need a single pass
to build the go toolchain.

Signed-off-by: Anisse Astier <anisse@astier.eu>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/go: fixing host go cross compilation

Actually if GOHOSTOS == GOOS || GOHOSTARCH == GOARCH the go build system assume
it's not cross compiling and uses the same toolchain for both the host and the
target. This commit adds a patch to enable the explicit
GO_ASSUME_CROSSCOMPILING in go build system and updates to go package
accordingly.

Fixes:
http://autobuild.buildroot.net/results/3636b1ac5756a782fd7578186508aaf9d105e3e9/
http://autobuild.buildroot.net/results/25790dca7e19527bb360d7dfb325cd9cfc3b56cc/
and many more.

References:
https://github.com/golang/go/issues/25177
https://golang.org/cl/112156

Signed-off-by: Angelo Compagnucci <angelo@amarulasolutions.com>
Signed-off-by: Anisse Astier <anisse@astier.eu>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

docs/manual: using a branch name as FOO_VERSION does not work

For various reasons, we've always suggested users to avoid using a
branch as version string for their packages, because it does not work
as a they would expect:

  - it is not reproducible, because the branch may change between two
    builds that are done at different times;

  - it does not even follow the branch, as Buildroot anyway generates
    a local tarball, which it will reuse on subsequent builds.
    Furthermore, since we fetch and not pull, any existing local branch
    is not updated.

Yet, until recently, using a branch name would just work (with the
above limitations): the git tree was cloned, the branch checked out,
and the tarball created.

But with the advent of the git caching, using a branch name does not
work anymore. Indeed, we now do a git-fetch, and that does not create
a local master branch. So we can't check out master, because it does
not exist locally. And for other branches, as noticed above, the local
branch does not get udpated to the remote one.

Furthermore, the local branches are only created by chance, again as a
side-effect of trying to fetch the "special refs".

So, we can't say that we reliably support the use of a branch name.

Update the manual to state that using a branch does not work. Remove
the 'stable' example, as it looked like the name of a stable branch;
instead, replace it with a version string that ressemble a tag.

Fix the layout of the manual by making the version examples an actual
bulleted list.

Note: the above is only entirely true for git. For Mercurial, CVS and
subversion, the status may be mixed, but nonetheless, using branches is
still a bad idea, if at least because it is not reproducible, and
because Buildroot does not even follow the branch. So, we do not
differentiate between the various SCMs, and just flatly state that using
a branch name is not supported.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

strace: disable (again) mpers for musl

It turns out that mpers support in strace still does not play nicely
with musl libc. As explained in commit 7892a778b (package/strace:
disable libmpers with musl toolchains) headers mixup causes gcc header
to be included, instead of the musl one, resulting in conflicting types.
Commit 1088372941 (strace: bump to version 4.21) incorrectly enabled
mpers for musl. Revert that.

Fixes:
http://autobuild.buildroot.net/results/345/3452419498c074ca66f36f0f87263fa10662ac86/

Cc: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/dahdi-linux: use package's DL_DIR

Since we reworked the download infra, the location for packages to look
for their files has moved to a per-package directory.

For systems where a download of the dahdi firmware files was already
done in a version prior to the rework, all was working fine so far,
because the files were indeed in the main DL directory.

But for systems where the download is first attempted after the rework,
the files are not found (even though they are properly downloaded).

Fix the location where dahdi-linux looks for its extra files.

Reported-by: ***** ***** <zyama.abel@mail.ru>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: ***** ***** <zyama.abel@mail.ru>
Cc: Carlos Santos <casantos@datacom.ind.br>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/asterisk: use package's DL_DIR

Since we reworked the download infra, the location for packages to look
for their files has moved to a per-package directory.

For systems where a download of the asterisk sound files was already
done in a version prior to the rework, all was working fine so far,
because the files were indeed in the main DL directory.

But for systems where the download is first attempted after the rework,
the files are not found (even though they are properly downloaded).

Fix the location where asterisk looks for its extra files.

Reported-by: ***** ***** <zyama.abel@mail.ru>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: ***** ***** <zyama.abel@mail.ru>
Cc: Carlos Santos <casantos@datacom.ind.br>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

support/dependencies: check that PATH does not contain CWD

A person on IRC reported a build failure with the util-linux package,
looking like this:

for I in uname26 linux32 linux64        ; do \
cd /home/aep/consulting/chargery/tracker/output/target/usr/bin && ln -sf setarch $I ; \
done
[...]
/bin/sh: line 1: ./ln: cannot execute binary file: Exec format error
/bin/sh: line 1: ./ln: cannot execute binary file: Exec format error
/bin/sh: line 1: ./ln: cannot execute binary file: Exec format error

The issue was an empty path in the PATH variable, which means "current
working directory", causing a "ln" binary built by util-linux for the
target to be used instead of the system-provided "ln".

We already check a number of things in the PATH and LD_LIBRARY_PATH
variables in support/dependencies/dependencies.sh, but we were not
checking that PATH did not contain an empty path.

This commit fixes that and takes this opportunity to simplify the test
code for PATH and LD_LIBRARY_PATH.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[Thomas: improve commit log.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

docs/website: update news.html for 2018.05-rc1

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

Update for 2018.05-rc1

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

CHANGES: additional updates for 2018.05-rc1

 - Add some explanations about the Blackfin removal, as suggested by
   Yann E. Morin.

 - Add list of resolved issues.

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

CHANGES: update in preparation for 2018.05-rc1

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

package/mplayer: remove package

This package causes numerous build errors. During the last hackaton it
was decided to remove this package:
http://lists.busybox.net/pipermail/buildroot/2018-April/217514.html

Fixes
http://autobuild.buildroot.net/results/bda/bda461391c9e02e313fda887b75099e619527a7a/

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

package/tovid: switch from mplayer to mpv

The mplayer package will be removed from buildroot, tovid can use mpv
instead. Support for mpv was added upstream:
https://github.com/tovid-suite/tovid/commit/05cbce2ea130e50f6bf1599816dd0d7bf3ace91a

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

package/tovid: bump version

Support for mpv is not present in the tarball version 0.35.2, therefore
we bump this package to current git master branch to prepare the switch
from mplayer to mpv.

Added license hash.

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

package/libplayer: remove mplayer support

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

Revert "libcrossguid: bump to version 0.2.2"

This reverts commit dededf0ff2177e5aeb051db25e178bb951debce5.

Building the only package depending on libcrossguid, Kodi, breaks due to
this bump:

/home/buildroot/br4/output/build/kodi-17.6-Krypton/xbmc/utils/
StringUtils.cpp: In static member function ‘static std::__cxx11::string
StringUtils::CreateUUID()’:
/home/buildroot/br4/output/build/kodi-17.6-Krypton/xbmc/utils/
StringUtils.cpp:1194:10: error: ‘GuidGenerator’ does not name a type
   static GuidGenerator guidGenerator;
          ^~~~~~~~~~~~~
/home/buildroot/br4/output/build/kodi-17.6-Krypton/xbmc/utils/
StringUtils.cpp:1195:15: error: ‘guidGenerator’ was not declared in this
scope
   auto guid = guidGenerator.newGuid();
               ^~~~~~~~~~~~~

This happens due to changes in libcrossguid:
https://github.com/graeme-hill/crossguid/blob/master/README.md

"This is version 0.2 of CrossGuid. If you all already using CrossGuid and
your code uses GuidGenerator then you are using version 0.1."

Even Kodi master branch has no support for libcrossguid-0.2.x yet so we
revert the bump.

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

uclibc: fix ldconfig for x86_64

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

log4cplus: bump to version 2.0.0

- LICENSE has been updated to add threadpool license (BSD-like)
- Add hash for license file

Fixes:
 - http://autobuild.buildroot.net/results/a4690fd3e46354776f6d6addd17226622c5f4961

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

package/znc: bump version to 1.7.0

Increased minimum gcc version to 4.8 as mentioned in the Changelog:
https://wiki.znc.in/ChangeLog/1.7.0

Removed all patches:
- 0001 was applied upstream:
https://github.com/znc/znc/commit/483074cbf82a48a3ff01838fcbaec00553d1882c

- 0002 is not needed anymore after upstream switched Csocket to a git
  submodule:
https://github.com/znc/znc/commit/9166e6ed40f0c7a85eee0d024b12510c13863f55#diff-8a725997e024d92d67b33a90cffb8342

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

libkrb5: bump to version 1.16.1

Update hash for license file (update to year)

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

python-remi: bump to version 1.1

Signed-off-by: Gwenhael Goavec-Merou <gwenhael.goavec-merou@trabucayre.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

minissdpd: bump to version 1.5.20180223

Update hash of license file (update in year)

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

package/apr-util: add optional support for openldap

apr-util contains optional support for openldap:

$ ls -1 output/target/usr/lib/apr-util-1/apr_ldap*
output/target/usr/lib/apr-util-1/apr_ldap-1.so
output/target/usr/lib/apr-util-1/apr_ldap.la
output/target/usr/lib/apr-util-1/apr_ldap.so

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

package/apr-util: add optional support for postgresql

apr-util contains optional support for postgresql:

$ ls -1 output/target/usr/lib/apr-util-1/apr_dbd_pg*
output/target/usr/lib/apr-util-1/apr_dbd_pgsql-1.so
output/target/usr/lib/apr-util-1/apr_dbd_pgsql.la
output/target/usr/lib/apr-util-1/apr_dbd_pgsql.so

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

qemu_arm_vexpress: Linux kernel update to 4.16.7

Update qemu_arm_vexpress target to Linux kernel 4.16.7.

Signed-off-by: Florian La Roche <Florian.LaRoche@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

libkcapi: split BR2_PACKAGE_LIBKCAPI_APPS

Since commit https://github.com/smuellerDD/libkcapi/commit/af313accf18887b4f6dbd7306c2389ddf271c9b7,
hasher test application depends on dlfcn.h. To avoid putting all apps
under !BR2_STATIC_LIBS dependency, split BR2_PACKAGE_LIBKCAPI_APPS into
several variables.

Moreover, BR2_USE_MMU is only a dependency of BR2_PACKAGE_LIBKCAPI_TEST
as other applications don't use fork.

Finally, enc or dgst applications could be enabled in a next patch if
needed.

Fixes:
 - http://autobuild.buildroot.net/results/7c57f3025030eff41a5cde52759821249859caf1

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

python-cython: new package

Signed-off-by: Joseph Kogut <joseph.kogut@gmail.com>
Reviewed-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

dropbear: Do not build static binary

Dropbear 2018.76 now uses the --enable-static option to indicate that a static
binary should be built. This will incorrectly pick up the generic buildroot
option intended for building static libraries, causing an unwanted static
binary build with BR2_SHARED_STATIC_LIBS.

Fix by appending an --disable-static configure flag, overriding the buildroot
default.

Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Reviewed-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

configs/mx53loco: Bump kernel and U-Boot versions

Bump kernel to version 4.16.7 and U-Boot to 2018.05.

Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

busybox: S01mdev: fix module autoloading

Commit b4fc5a180c (package/busybox: support spaces in module aliases in
mdev) changed the mdev coldplugging to handle sysfs path elements and
modalias values containing spaces.  This unfortunately doesn't work as was
recently reported:

http://lists.busybox.net/pipermail/buildroot/2018-May/220903.html

The problem is that sort -z also expects the fields of the input files to be
zero terminated, which is not the case for modalias sysfs entries.

So drop the -z option to sort.  Spaces in modalias entries could be handled
with the xargs -d '\n' option, but that is unfortunately not supported by
the busybox applet.  Instead, use tr to convert newlines to zeros so we can
use xargs -0.

Reported-by: Daniel Palmer <daniel@0x0f.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

mesa3d: fix build with gcc <= 4.7 on some architectures

The configure.ac check for atomic intrinsics assumes that if an atomic
intrinsic is not available, it's because we must link with
libatomic. Except that libatomic is not always available, for example
with gcc <= 4.7.

To fix this, this commit adds a patch that reworks the atomic check in
configure.ac. It has been submitted upstream, and tested in a number
of different combinations.

Fixes:

  http://autobuild.buildroot.net/results/b045488d880a4013ed82937e7115b7636867e01d/

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

zeromq: check if -latomic is needed

Fixes:
http://autobuild.buildroot.net/results/b149aa2ee00e4d6a53c884cf99ecb2dd8af58b65/
http://autobuild.buildroot.net/results/e4b3616ac2695d3b6898185a70da6509b1faa2b8/

Patch the package to check if -latomic is needed to be added, depending on the
result of AC_LINK_IFELSE.

The patch was sent to upstream, see:
https://github.com/zeromq/libzmq/pull/3083

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

package/{mesa3d, mesa3d-headers}: bump version to 18.0.3

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

sudo: bump to version 1.8.23

Use += for the SUDO_CONF_ENV conditional assignment to avoid override of
SUDO_CONF_ENV assignments that might be added in the future.

Add license file hash.

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

python-raven: bump to version 6.7.0

Signed-off-by: Joseph Kogut <joseph.kogut@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

lxc: fix build with static libcap and shared gnutls

Fixes:
 - http://autobuild.buildroot.net/results/b655d6853c25a195df28d91512b3ffb6c654fc90

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[Thomas:
 - add Upstream-status
 - use proper numbering]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

uclibc: fix clock_nanosleep declaration

Fixes autobuild errors:
 http://autobuild.buildroot.net/results/a58007263355e2734a8074d52b1b26b88973c39e
 http://autobuild.buildroot.net/results/7d6fd0872efd23c7de552ab637956902b43c3f58

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Reported-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gst1-plugins-ugly: fix x264 compile failure

Add upstream patch to fix compile with latest x264 library.

Fixes [1]:

  gstx264enc.c: In function 'plugin_init':
  gstx264enc.c:2900:36: error: 'x264_bit_depth' undeclared (first use in this function); did you mean 'x264_picture_t'?
     default_vtable.x264_bit_depth = &x264_bit_depth;
                                      ^~~~~~~~~~~~~~
                                      x264_picture_t

[1] http://autobuild.buildroot.net/results/a766a28c584534b6a30839cfd98428d840bce3f2

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libfreeimage: correct 0005-Manage-powf64-with-glibc.patch line endings

Fixes:
http://autobuild.buildroot.net/results/131/131f97d2c3525794163e5c7863bb8b8523fb05ba/

The libfreeimage source code uses DOS line endings, which gets stripped by
patchwork so the patch no longer applies.  Fix it by converting line endings
of the patch hunks to DOS (but not the patch meta data).

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

binutils: replace hard-links with soft-links to fix rpath

binutils installs its binaries both as bin/<tuple>-<tool> and as
<tuple>/bin/<tool>, and hardlinks are used to reduce disk space
consumption. This causes a problem for host-binutils with our rpath
fixing logic done by "make sdk".

Indeed, the fix-rpath script starts by fixing up the rpath of
bin/<tuple>-<tool>, and sets the RPATH to $ORIGIN/../lib/. Then
fix-rpath moves on to <tuple>/bin/<tool>, and doesn't find the library
the tool depends on, and clears the RPATH. The result is that the
binutils tool are not usable.

Note that this is only visible currently on the ARC architecture,
because on this architecture, binutils is fetched from git, which
causes host-flex to be built, and some binutils tools to use the libfl
shared library. Therefore, the binutils tools don't use just the
standard C library (which is provided by the system) but also libfl
from $(HOST_DIR)/lib, and therefore if the RPATH isn't set correctly,
those tools don't work properly.

In order to address this, this comit adds a post-install hook to
host-binutils that replaces those hard links by symbolic links. It is
worth mentioning that library loading and RPATH usage occurs *after*
resolving the symbolic links, which makes this solution work.

Fixes:

  http://autobuild.buildroot.net/results/b2562b05d397d4e1ffe0f8d2f4ce4c84ab6feae1/

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

udftools: fix static linking against readline

One program of udftools uses the readline library, but fails to build
in static linking configurations. In order to fix this, we teach
udftools configure.ac to use pkg-config to detect the readline
library, and to use the proper flags provided by pkg-config.

This obviously requires an autoreconf of the package, and the addition
of host-pkgconf in the dependencies.

Fixes:

  http://autobuild.buildroot.net/results/113a94049d89b8f065112e5d4482667a7b7fb843/

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

readline: install a .pc file

Some packages (such as udftools) fail to build against readline in
static linking configurations. To fix this, we install the .pc file
provided by readline.

Unfortunately, this .pc file is by default not correct, so a patch is
added to fix it, and this patch requires to re-autoconf the package
(but not autoreconf since only autoconf is used).

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

udftool: add missing dependency on readline

udftool/Config.in selects BR2_PACKAGE_READLINE, but udftool does not
depend on it. Readline is not a runtime dependency: one of the udftool
program links against it.

This commit fixes this inconsistency.

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

.gitlab-ci.yml: update after removal of freescale defconfigs

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

linux: ensure correct make targets are used for cuImage/simpleImage.<dtb>

Fixes https://gitlab.com/buildroot.org/buildroot/-/jobs/66561794

LINUX_DTS_NAME may end up with a leading space because of the += logic, and
may contain multiple dts files - Neither of which works when we construct
the {cu,simple}Image.$(LINUX_DTS_NAME) make target name.

Fix it by using the first word in the variable.

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

libmhash: add hash for license file

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

libmnl: add hash for license file

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

libmng: add hash for license file

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

libnetfilter_acct: add hash for license file

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

libmemcached: add hash for license file

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

package/Config.in: fix alphabetical order

Signed-off-by: Jerzy Grzegorek <jerzy.m.grzegorek@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/patch: fix xattr option

The name of the configure option is, and has always been,
 --enable-xattr, not --enable-attr. Otherwise, configure
whines:
    configure: WARNING: unrecognized options: --enable-attr

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

libupnpp: fix build with uclibc

Add patch to avoid failing with error if host_os doesn't start with
linux

Fixes:
 - http://autobuild.buildroot.net/results/cfb332a169863bcb3e6bc6ee7d6f6199f856b0b8
 - http://autobuild.buildroot.net/results/2f6f464f6360ed9ea4c238f503d2c3b8ab3cbd86

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

log4cplus: fix license

Some files of log4cplus are licensed under BSD-2-Clause as stated in
LICENSE file: "Each file of log4cplus source is licensed using either
two clause BSD license or Apache license 2.0."

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

can-utils: bump to version 2018.02.0

Disable for musl. can-utils now uses the error() glibc extension.

Use Makefile as license file.

Change homepage link to the github repo. The README.md file is much more
informative than the elinux.org wiki page.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ebtables: drop obsolete comment

Commit 30c4b5843f (ebtables: enable for musl) enabled musl build, but
forgot to remove the comment text. Fix that.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/freeswitch: add optional support for postgresql

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

libnice: bump to version 0.1.14

- Remove patch (already in version)
- Add gnutls mandatory dependency
- Add hash for license files

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

lcms2: bump to version 2.9

- Remove two patches (already in version)
- Add hash for license file

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

ghostscript: don't use lcms2art

Delete lcsm2art directory to use the buildroot lcms2 library

Fixes:
 - http://autobuild.buildroot.net/results/cda9c22bf29278cc24ab852094df19b773d0f151

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

libmpeg2: add hash for license file

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

libnpth: add hash for license file

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

libnatpmp: add hash for license file

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

libndp: bump to version 1.6

Add hash for license file

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

libmcrypt: add hash for license file

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

libmad: add hash for license file

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

libmbus: add hash for license file

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

libkrb5: add dependency on host-bison

libkrb5 needs yacc

Fixes:
 - http://autobuild.buildroot.net/results/0bd0011978ecf0958ad459ae1aace564f4dde870

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

configs/freescale_imx6ulevk: Remove target

Currently this target fails to build:

./scripts/dtc-version.sh: line 17: dtc: command not found
./scripts/dtc-version.sh: line 18: dtc: command not found
*** Your dtc is too old, please upgrade to dtc 1.4 or newer

Even though the fix for this build failure is easy, I prefer to
maintain only imx6ulevk_defconfig, which is a target for the
same board, but uses mainline kernel and U-Boot versions instead.

Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/66561695
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/freescale_imx31_3stack: Remove target

Currently this target fails to build:

arch/arm/mach-imx/built-in.o: In function `init_mmdc_lpddr2_settings':
platform-spi_imx.c:(.text+0x8f18): undefined reference to `imx6sll_lpddr2_freq_change'
arch/arm/mach-imx/built-in.o: In function `imx6ul_cpuidle_init':
platform-spi_imx.c:(.init.text+0x195c): undefined reference to `mx6ull_lpm_wfi_end'
platform-spi_imx.c:(.init.text+0x1960): undefined reference to `mx6ull_lpm_wfi_start'
platform-spi_imx.c:(.init.text+0x1964): undefined reference to `imx6ull_low_power_idle'
arch/arm/mach-imx/built-in.o: In function `imx6sl_init_late':
platform-spi_imx.c:(.init.text+0x56f0): undefined reference to `imx6sll_cpuidle_init'
Makefile:969: recipe for target 'vmlinux' failed

NXP 4.9 kernel does not properly support mx31, so better remove this target.

mx31_3stack board is currently supported in U-Boot and kernel mainline,
so if someone is interested on this board, then a target can be introduced
later which uses these mainline components.

Currently I do not have access to this board, so remove it its defconfig
for now.

Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/66561688
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/freescale_imx6sololiteevk: Remove target

Currently this target fails to build:

Error: No BOOT_FROM tag in board/freescale/mx6slevk/imximage.cfg.cfgtmp
arch/arm/imx-common/Makefile:91: recipe for target 'u-boot-dtb.imx' failed

This problem has been already solved in upstream U-Boot, but not
on NXP version of U-Boot.

To avoid the maintainance burden of keeping two defconfigs for the
same board, remove freescale_imx6sololiteevk_defconfig in favor
of imx6slevk_defconfig, which uses U-Boot and kernel mainline.

Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/66561693
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/Config.in: fix alphabetical order

Signed-off-by: Jerzy Grzegorek <jerzy.m.grzegorek@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

nasm: Fix compilation with GCC 8

host-nasm fails to compile with GCC 8:

In file included from ./include/nasm.h:46,
                 from asm/nasm.c:47:
./include/nasmlib.h:194:1: error: ‘pure’ attribute on function returning ‘void’ [-Werror=attributes]
 void pure_func seg_init(void);
 ^~~~

Pull a patch from upstream that removes the stale declaration of
seg_init, which was eliminated a long time ago.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dropbear: Fix host key loading with 521 bit ecdsa keys

Dropbear 2018.76 changed the default ecdsa host key size form 521 to 256
bits, but this breaks systems with an existing 521 bit key, blocking ssh
logins.

Apply the upstream fix from https://secure.ucc.asn.au/hg/dropbear/rev/0dc3103a5900 :

  Only advertise a single server ecdsa key when -R (generate as required) is
  specified. Fixes -R now that default ecdsa key size has changed.

[Peter: apply-patches.sh does not like suffix-less filename, so include
patch in Buildroot]
Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

docs/website: update for 2018.02.2

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

gnupg2: bump to version 2.2.7

Cc: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

libgpg-error: bump to version 1.31

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

libgpgme: bump to version 1.11.1

Drop upstream patch.

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

openvpn: bump to version 2.4.6

Note that CVE-2018-9336 fixed in this version does not affect Buildroot
since it is Windows specific.

Drop interoperability with older busybox versions. We removed support
for busybox version selection long ago.

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

liblogging: bump to version 1.0.6

Add hash for license file

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

liblog4c-localtime: add hash for license file

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

package/bash: use target finalize hook

Arnout reported an issue with the upcoming top-level parallel build.
The file /etc/shells can be modified by several packages (shell providers)

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

package/bash: add missing double-dollar for /etc/shells

Commit 4d279697afbf8fb295274784103be2b837113d5e missed one comment
from Yann E. Morin about the double-dollar.

Signed-off-by: Romain Naour <romain.naour@smile.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>