pixman: disable vmx for powerpc CS toolchains

The ppc CS toolchains have flaky VMX (AKA altivec) implementation
causing build failure. Disable it for these toolchains. Fixes:
http://autobuild.buildroot.net/results/3b4/3b4cda40f074578f6b35b1fb381ba37b93174e86/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

p11-kit: bump to version 0.23.2

0002-rpc-transport.c-include-sys-select.h-for-fd_set.patch is upstream
so drop it.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/xfsprogs: disable crc32selftest

xfsprogs generates a crc32 table, and wants to test it. So it builds a
native crc32selftest binary. So far, so good; except the host may not
have the required headers (uuid/uuid.h), and probably libs (like
libuuid).

And anyway, testing the table on the host is flawed, as it does not mean
it would be correct on the target (due to differences in bitness,
endianness, alignment...).

So we just disable the crc32selftest for Buildroot.

Fixes:
    http://autobuild.buildroot.org/results/5d0/5d0ed24b4d368fed100bc3c4959520ebaa48e693/
    http://autobuild.buildroot.org/results/aef/aef223ee2646599eaab7c82f63ff5f19569d69a5/
    http://autobuild.buildroot.org/results/fb6/fb6679f3b1473169702c303ee5060abe7559b0b3/
    ....

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Use libgudev as regular dependency

Since eudev 3.1.3 libgudev is not shipped with this package,
so libgudev can be used as regular dependency regardless of
udev implementation selected.

Fixes following build error:
  http://autobuild.buildroot.net/results/320/320c052bda0f1b5afb1e5c83a7fb4dca6227c5a1//
  http://autobuild.buildroot.net/results/3b0/3b096f882030325b9290ad8860bd9cd373a11dc2//

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libgudev: remove systemd dependency

Since version 3.1.3 eudev doesn't provide internal libgudev,
so this package can now be used as an ordinary stand-alone
package.

[Peter: add BR2_PACKAGE_HAS_UDEV dependency]
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

nano: bump to version 2.5.0

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gdb: bump 7.10 series to 7.10.1

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

lvm2: bump version to 2.02.137

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

lighttpd: bump version to 1.4.38

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Reviewed-by: Alexander Dahl <post@lespocky.de>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcr: needs host-intltool

Fixes:
http://autobuild.buildroot.net/results/c31/c3177224ca2488cf4740a2d22d1a861872d2dcde/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/mpython-m2crypto: bump version

Fixes:

http://autobuild.buildroot.org/results/ee1/ee1bb80214a51c212712de8193b89a32e9901408/
http://autobuild.buildroot.org/results/ce8/ce8d5d68eef312ca168f351b845920eb62ca4eb1/
http://autobuild.buildroot.org/results/500/5004a2a9cabe75669bbe90f3ee35ab3869617fc9/

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

package/kodi: fix fallout after infozip rename

Commit f0c9470 (infozip: rename to zip) forgot to propagate the rename
down to the only user of the host variant of infozip.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Acked-by: Luca Ceresoli <luca@lucaceresoli.net>
Acked-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcc: bump 5.x series to version 5.3.0

201-libgcc-remove-unistd-header.patch is upstream so remove it.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Tested-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcc: host-cloog is not needed for graphite support on gcc 5.x

See https://gcc.gnu.org/gcc-5/changes.html

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

flite: fix alsa version check

The alsa version check broke with the recent alsa-lib bump to 1.1.0 as it
was only checking subminor >= 11.

Fixes:

http://autobuild.buildroot.net/results/b62/b62cd6de04ff83bf7025a8f1aca49dbcbc8d7e2d/
http://autobuild.buildroot.net/results/2c9/2c98458fef8c38ff36cd20d77c2d651129146857/
http://autobuild.buildroot.net/results/ca7/ca78c1029ef1d91b66bbc91c5e5269bf663afe8e/
http://autobuild.buildroot.net/results/eac/eacb973e7fcce385d1686ab364f60dec9e0dd2e9/

And many more.

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

docs: website: update for 2015.11.1

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

package/pixman: disable tests

Tests are broken on some architectures because of an incomplete fenv.

Since the tests are not really useful in the Buildroot environment, just
disable them.

Fixes a lot of build failures:
    http://autobuild.buildroot.org/results/cd7/cd722620832c28d22eacfc43336f6952b82305ff/
    http://autobuild.buildroot.org/results/3f4/3f4f1e88988d66b052432885b4f044a0eddcee2c/
    ...

Pixman was already autoreconfed (but that was not needed). Since we now
really need to autoreconf it, add a comment stating why we do.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

boost: add target build step

The build step for the target was missing, causing the compilation to happen
during the staging install step.  Add a dedicated build step like we do for
the host build.

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

boost: reorder build steps

The host and target build steps were mixed between eachother. Reorder them
so we have the target steps before the host ones.

No functional change, but helps readability.

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

boost: make package less verbose when building with -s

The boost build is very verbose (11K+ lines even when no sub options are
enabled).  While this can be handy for debugging, make is less verbose when
'make -s' is used, similar to how we do it for autotools.

With this, make -s boost outputs less than 100 lines.

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

at91bootstrap3: bump to v3.8

In preparation for sama5d2 xplained support, bump at91bootstrap3 to 3.8

[Peter: add .hash file]
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libpng: security bump to version 1.6.20

Fixes:
CVE-2015-8126 - incorrect implementation of png_set_PLTE() that uses
png_ptr not info_ptr, that left png_set_PLTE() open to this vuln.

(fix in previous release was incomplete)

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

bdwgc: fix compile for sparc

We need to use a patch from master git.
https://github.com/ivmai/bdwgc/commit/e2bf29b7801357c1ad28c2948a3f88c655dd722a

Fixes following autobuilder failures:
http://autobuild.buildroot.net/results/220b3d49b8ad891e5dd3aae759319100a1500701/
http://autobuild.buildroot.net/results/b7465d175f03c9461693a92885028723772c7814/

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libatomic_ops: enable sparc64 / sparcv9

Works fine for sparc64.

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gnutls: needs argp-standalone for musl/uclibc

It's required for crywrap (tools), fixes:
http://autobuild.buildroot.net/results/2d9/2d98fced7ff7c3b5c39a97eb1c8db7dd651fa86c/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

directfb: disable for sparc

Requires atomic operations.
Fixes following autobuild failure:
http://autobuild.buildroot.net/results/8205100c166659c47daff66b1e63edc82eeb7930/

[Peter: also hide toolchain comment]
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/qemu: specify drive image file format (raw)

Specify that the drive image file format is raw to avoid warnings and
restrictions.
Also switch to -drive syntax for all configs.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

fakeroot: disable capabilities

These don't autodetect consistently and are process rather than file
capabilities, so not very useful in the buildroot context.
Otherwise we'd have to unconditionally pull host-libcap (or fix the
detection, but again, not useful so there's no need).

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libmicrohttpd: bump to version 0.9.47

Experimental SPDY support was dropped so drop the configure option.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wireshark: add optional geoip support

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wireshark: add optional gnutls support

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wireshark: gcrypt support is optional

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wireshark: add optional c-ares support

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wireshark: add optional libnl suport

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wireshark: enable GUI options

Enable gtk2, gtk3 & qt5 automatic GUI options.
gtk2 & qt5 only build-time tested, gtk3 runtime-tested with wayland.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wireshark: needs host-pkgconf

Needed for detection of various optional components.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wireshark: forcibly disable lua

If host arch = target arch and liblua is in the host the buildsystem
will pick it up and try to use it since it passes basic build tests.
Forcibly disable it, since it causes no build failure, rather runtime
failures because said liblua is not present on the target.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

python-lxml: bump to 3.5.0

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openssl: security bump to version 1.0.2e

Fixes:
CVE-2015-3193 - BN_mod_exp may produce incorrect results on x86_64
CVE-2015-3194 - Certificate verify crash with missing PSS parameter
CVE-2015-3195 - X509_ATTRIBUTE memory leak

Enable IDEA as well since otherwise the build breaks (always great
upstream) - it's no longer patent encumbered.

[Peter: correct sha256]
Signed-off-by: Gustavo Zacarias <gustavo.zacarias@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

connman: disable for sparc

Fixes:
http://autobuild.buildroot.net/results/333/3336df777c758cba9a660a0f6f31646258baec52/
http://autobuild.buildroot.net/results/efe/efeb197654b552ad9a80e102226c3e507640f037/

Connman uses __sync_fetch_and_add, which is not available for sparc.

[Peter: add !sparc dependency to comment, add autobuild reference]
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

stunnel: bump to version 5.27

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dropbear: bump to version 2015.71

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gesftpserver: bump to version 0.2.1

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

hostapd: make ACS optional

Make ACS (Automatic Channel Selection) support in hostapd optional.
There's a clash of standard vs. propietary solutions, hence when hostapd
has builtin ACS support (standard) and it's enabled (channel=0 in the
config) this will preclude the propietary driver from using ACS (it
won't work at all).

See:
http://e2e.ti.com/support/wireless_connectivity/f/307/p/465333/1669270

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

cmake: bump to version 3.4.1

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

sysstat: bump version to 11.0.8

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

zsh: bump to version 5.2

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Makefile: fix copyright year

It is no longer 2014.

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

busybox: add another upstream patch to fix (g)unzip

0002-unzip.patch was added in 69516e0 to fix a segmentation fault in the
gunzip applet.  However, it introduced a new issue that made the unzipping
of some files fail.

Add an upstream patch that fixes this new issue.

Fixes #8501.

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Jason Rush <rush0033@hotmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/gawk: do not use readeline ormpfr in host variant

gawk has an optional dependency on mpfr (and thus gmp) and readline,
and will probe for them. If they are present, they are used; if they
are missing, that's not an error. mpfr (and gmp) is used for "BIGNUM"
support on gawk; readline is used by the gawk debugger.

However, mpfr (bringing gmp) are also host-packages in Buildroot, but in
the standard build order (i.e. a plain 'make'), they are built after
gawk. Ditto readline (from ncurses).

If the user has the development files for gmp and mpfr, then gawk is
linked to them. Ditto readline.

Now, further on in the build, we build gmp and mpfr (for gcci or guile),
so we install them in the host dir. Ditto readline (for gdb, ncurses
itself and a few other packages...)

But because we forcibly set an RPATH tag on all our host binaries, our
host gawk will now dynamically link with our versions, when it was in
fact built against the host ones.

This did not seem to cause any harm so far, but is far from ideal.

Since we do not really need BIGNUM or the debugger in our host gawk, we
just forcibly disable them and configure gawk without readline or mpfr
(there's no switch for gmp, but it's not a direct dependency, it comes
just with mpfr).

[Adjust comment as suggested by Thomas/Yann]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Peter Korsgaard <jacmet@uclibc.org>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

support/check-host-rpath: remove trailing '/' in host dir

Despite the comment saying so, the trailing '/' in the host directory is
not removed. Note however that it is properly removed from extracted
RPATH tags.

This is not visible when the host directory is our default $(O)/host
location, but breaks for user-supplied external host directory, when
the user leaves a trailing slash in the path.

Fix that.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Peter Korsgaard <jacmet@uclibc.org>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

linux-headers: bump 3.{2, 12}.x series

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

p11-kit: bump to version 0.23.1

It's required for the new gnutls version, fixes:
http://autobuild.buildroot.net/results/3ec/3ecac965455dc169162cbdb50f8ff07e2880e504/

Patch status: reported upstream:
https://bugs.freedesktop.org/show_bug.cgi?id=93211

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

rrdtool: add patch to disable documentation

Add patch to disable documentation. Fixes:
http://autobuild.buildroot.net/results/6fb/6fb94127af6ff53bdcf52ee0eec8971ccd8858f5/

Patch status: pull req https://github.com/oetiker/rrdtool-1.x/pull/690

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

fakeroot: bump to version 1.20.2

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libhttpparser: add a hash file

Signed-off-by: Renaud AUBIN <root@renaud.io>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dbus: bump to version 1.10.6

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libcurl: bump version to 7.46.0

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Merge branch 'next'

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

docs/website/news.html: add 2015.11 announcement link

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

Kickoff 2016.02 cycle

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

gnutls: bump to version 3.4.7

The 3.4 series has been promoted to stable.
Handle autodeps more concisely (idn, p11-kit, zlib).
libtasn1 is now mandatory, since otherwise gnutls uses the bundled
version it makes no sense to try that because of target duplication.
Disable tpm support since we've got no trousers package.
Disable libdane support since we've got no dane package.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libhttpparser: bump to 2.6.0

The http-parser repository has been transfered from joyent to nodejs.
Bump version to 2.6.0.

Signed-off-by: Renaud AUBIN <root@renaud.io>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

openldap: bump version to 2.4.43

- Bump version to 2.4.43
- Use md5 and sha1 upstream provided hashes. This combination is
  stronger than a sha256 hash.
- Switch to an http download site to avoid problems with firewalls
  blocking the ftp.

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

cryptodev-linux: bump to version 1.8

Patches are upstream so remove them.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

openocd: bump to 0.9.0

Remove obsolete compilation error patch and update jimtcl related patch.

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

dhcpcd: bump to version 6.9.4

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

samba4: bump to version 4.3.2

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/nss-pam-ldapd: new package

Name Service Switch (NSS) module that allows your LDAP server to provide
user account, group, host name, alias, netgroup, and basically any other
information that you would normally get from /etc flat files or NIS. It
also provides a Pluggable Authentication Module (PAM) to do
authentication to an LDAP server.

[Thomas:
 - bump to version 0.9.6
 - use --disable-<foo> rather than --enable-<foo>=no.
 - fix license information: there is no LICENSE file, it is named
   COPYING.]

Signed-off-by: Doug Kehn <rdkehn@yahoo.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

miraclecast: new package

[Thomas:
  - add hash file
  - fix license information: use comma as a separator in
    <pkg>_LICENSE, and htable is under LGPLv2.1+ and not GPLv2.1+
    (which doesn't exist)]

Signed-off-by: Ryan Barnett <ryanbarnett3@gmail.com>
CC: Baruch Siach <baruch@tkos.co.il>
CC: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Update for 2015.11

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

lvm2: bump version to 2.02.136

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

conntrack-tools: work around build issue with musl

Building conntrack-tools with kernel headers >= 4.2 + musl fails due to
a well-known symbol clash that occurs when userspace and kernel headers
are included simultaneously (see [1], question 7, for details).

In the case of conntrack-tools, the inclusion of both 'netinet/in.h' and
'linux/in.h' occurs inside the C helper files (src/helpers/*.c)
indirectly via e.g. 'libnetfilter_conntrack/libnetfilter_conntrack.h',
which itself includes 'netinet/in.h', and 'linux/netfilter.h', which
includes 'linux/in.h' in kernel headers >= 4.2.

The approach to solving this type of conflict with musl usually involves
removing the inclusion of kernel headers or refactoring the code so as
to avoid the mentioned simultaneous inclusion. This is unfortunately
non-trivial in the case of conntrack-tools since the clashing headers
get included indirectly by headers that are strictly necessary (because
of definitions used in some helper callbacks).

Work around the issue by defining __GLIBC__ when musl is used. This
eliminates the conflicts as the kernel headers avoid redefining certain
symbols when they see __GLIBC__ defined (linux/libc-compat.h). Note that
other glibc-compatible libraries, like uClibc, already do that
internally.

Fixes:
  http://autobuild.buildroot.net/results/66e/66ec247fa0fc385bef8d2084c65bf5cad3a8e8ca/
  http://autobuild.buildroot.net/results/624/624a0d48decd819eb58cbb3c58ee904b87ebfb21/

[1] http://wiki.musl-libc.org/wiki/FAQ

Signed-off-by: Rodrigo Rebello <rprebello@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libsigsegv: fix sparc64 compile

Compile tested for sparc/glibc and sparc64/glibc.
Fixes following autobuilder problem:
http://autobuild.buildroot.net/results/1d8e0c22594aef53f7a4de3491454aa78311c815/

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/libv4l: v4l-utils need locale

To reproduce the setlocale-related build error use this minimal
defconfig:

BR2_TOOLCHAIN_BUILDROOT_CXX=y
BR2_PACKAGE_LIBV4L=y
BR2_PACKAGE_LIBV4L_UTILS=y

After enabling locale support the build error is fixed:

BR2_TOOLCHAIN_BUILDROOT_LOCALE=y
BR2_TOOLCHAIN_BUILDROOT_CXX=y
BR2_PACKAGE_LIBV4L=y
BR2_PACKAGE_LIBV4L_UTILS=y

Fixes
http://autobuild.buildroot.net/results/f2e/f2e4e34fb14ce355a6e92afc83024b0445f52513/
http://autobuild.buildroot.net/results/fe5/fe56063ddd2232fea7699f5117497beb3deca11e/
http://autobuild.buildroot.net/results/1be/1be55f096ddef2aa38fdba681ed97cf74905477a/

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

valgrind: add a patch to fix MIPS R6 by replacing ADDI with ADDIU

The ADDI instruction has been removed for MIPS R6, so valgrind fails to
build for this revision level because it uses that assembly
instruction. Using ADDIU instead fixes the problem.

The patch has been sent upstream:

  https://bugs.kde.org/show_bug.cgi?id=356112

Fixes:

  http://autobuild.buildroot.net/results/488/488937704b3b6bae8daed7b24da8c740b56f4d84/

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

imagemagick: bump version to 6.9.2-7

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

squid: bump to version 3.5.12

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

freetype: bump to version 2.6.2

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

harfbuzz: bump to version 1.1.2

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libglib2: force arm mode when building with thumb1

Fixes #8491

glib/valgrind.h contains inline asm not compatible with thumb1, so build it
in arm mode:

  CC       libglib_2_0_la-gstringchunk.lo
/tmp/ccSVEXME.s: Assembler messages:
/tmp/ccSVEXME.s:587: Error: shifts in CMP/MOV instructions are only supported
in unified syntax -- `mov r12,r12,ror#3'
/tmp/ccSVEXME.s:587: Error: shifts in CMP/MOV instructions are only supported
in unified syntax -- `mov r12,r12,ror#13'
/tmp/ccSVEXME.s:588: Error: shifts in CMP/MOV instructions are only supported
in unified syntax -- `mov r12,r12,ror#29'
/tmp/ccSVEXME.s:588: Error: shifts in CMP/MOV instructions are only supported
in unified syntax -- `mov r12,r12,ror#19'
/tmp/ccSVEXME.s:589: Error: lo register required -- `orr r10,r10,r10'
Makefile:2087: recipe for target 'libglib_2_0_la-gslice.lo' failed

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

core/printvars: allow dumping a set of variables

Dumping our 176164 variables can take quite some time (~12s here). What
takes the most time is sorting the variables (~9s), followed by the
parsing of our Makefiles (~3s), with the actual printing in the noise.

However, sometimes only one or a few variables are needed. For example,
one may want to retrieve the Linux build dir from a post-build hook (to
get the Linux' actual .config after our fixups and check for various
features).

Add the possibility to only dump the variables listed in $(VAR) which
must be passed as a make argument, like so:

    $ make -s printvars VARS="LINUX_DIR TOPDIR O"
    LINUX_DIR=/home/ymorin/dev/buildroot/O/build/linux-4.3 ($(BUILD_DIR)/$(LINUX_BASE_NAME))
    O=/home/ymorin/dev/buildroot/O/. (/home/ymorin/dev/buildroot/O/.)
    TOPDIR=/home/ymorin/dev/buildroot/buildroot (/home/ymorin/dev/buildroot/buildroot)

It is also possible to use make-appterns, like:

    $ make -s printvars VARS="BUSYBOX_%"

This is much faster (the time is just about the time it takes to parse
our Makefiles, 3s here) and easier to parse.

[Thomas: improve comment above the printvars target.]

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

support/download: fetch all refs on full git clone

When specifying BR2_LINUX_KERNEL_CUSTOM_REPO_VERSION, a user may want to
specify the SHA of a reference different than a branch or tag.

For instance, Gerrit stores the patchsets under refs/changes/xx/xxx, and
Github stores the pull requests under refs/pull/xxx/head.

When cloning a repository with --bare, you don't fetch these references.
This patch uses --mirror for a full clone, in order to give the user
access to all references of the Git repository.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: "Maxime Hadjinlian" <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

spi-tools: new package

This package contains some simple command line tools to help using Linux
spidev devices.

https://github.com/cpb-/spi-tools

The included version is based in a sha1 and not a tagged version as the
current master branch contains autotools support for build but there is
no tagged version with autotools support yet.

The only other quirk with the recipe is related to the VERSION define
which is normally obtained in build time via git for this package, but
that would fail within Buildroot as the package build is normally not
performed inside a git repository.
To work around that, I have added a hook which will substitute change
the VERSION to match that of the Buildroot package.

I have tested this package in an ARM target with a custom SPI device.

Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

squashfs: always enable xattr support

In 785fef10 xattr support was disabled, probably because the old version
did require the xattr package to be present to support it.
In a409d76e xattr support was made optional via (lib)attr, however the
squashfs package does in fact include all of the required bits to
support xattr without the need for the attr package.

So enable xattr support by default for both the host and target
variants.

Signed-off-by: Gustavo Zacarias <gustavo.zacarias@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libnet: PF_PACKET is always available on Linux

[Thomas: remove mistakenly added empty line.]

Signed-off-by: Joris Lijssens <joris.lijssens@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libnet: disable for musl toolchains

Signed-off-by: Joris Lijssens <joris.lijssens@gmail.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

at: rename patches so they are in a consecutive order

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

at: bump version to 3.1.16

- Bump version to 3.1.16
- Update the hash value
- Remove upstreamed patches:
  0001-configure.ac-remove-manual-compiler-check-with-AC_TR.patch
  0003-Makefile.in-add-LDFLAGS-to-linking-stage.patch
- Tweak patches for the new version:
  0004-Makefile.in-replace-IROOT-by-DESTDIR.patch
  0005-Makefile.in-make-install-fix.patch

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/zmqpp: fix missing install of static library

zmqpp's Makefile does not install the static library.

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/zmqpp: fix static build issues

The Makefile for zmqpp builds both shared and static libraries and the client
binary. This leads to several build issues in a pure static library context:
  * R_ARM_TLS_LE32 relocation not permitted in shared object
  * relocation R_ARC_32_ME against `_ZSt7nothrow' can not be used when making
    a shared object; recompile with -fPIC

We add a minimal patch to add some basic handling of building a shared or a
static library only.

Additionally, disable the client for static only builds as it depends on
building the shared library in zmqpp's Makefile. As there is already version
4.1.2 available which provides a CMake build file which solves this issue
(probably, not tested) we don't care for now.

Fixes:
http://autobuild.buildroot.net/results/345/345771eb488c60585e388fbbf4490df936e88e19/
http://autobuild.buildroot.net/results/21b/21b6912c70a5c300bdabde53bee6a1d9cc3bbb02/
http://autobuild.buildroot.net/results/d98/d9882d2ba00da16f76cea6d86a84cd4815ebbba2/

[Thomas:
 - don't change TARGET_CONFIGURE_OPTS, use ZMQPP_MAKE_OPTS instead.
 - simplify condition logic.]

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gpsd: work-around ICE when compiling for microblaze with GCC5

GCC 5.X generates an ICE when compiling gpsd for microblaze due
to optimizations, so lets disable it for now to works around
this problem.

A bug was reported to the gcc bug tracker:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68485

Signed-off-by: Sergio Prado <sergio.prado@e-labworks.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

uemacs: remove XCASE flag to allow musl builds

Fixes
http://autobuild.buildroot.net/results/e92/e929de08954684e275a2f682264551b6cb07886a/

Signed-off-by: Alex Suykov <alex.suykov@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/glmark2: Wayland-gl support also needs egl

We were missing the egl dependency when enabling Wayland GL support:
https://github.com/glmark2/glmark2/blob/master/src/wscript_build#L28

Fixes
http://autobuild.buildroot.net/results/858/8587d0cea86efa4c6e98817cadc7cce0b1b36e3d/
http://autobuild.buildroot.net/results/1f6/1f6853d5f7a83be60ccdf4f071a1df1491cf8137/
http://autobuild.buildroot.net/results/a4d/a4d830291ee8550b3c544ade7169218d5edd4d52/
http://autobuild.buildroot.net/results/536/536a8dc62783633943770a278dc9f1c6c5c0807c/

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

yad: needs gtk3-x11

Doesn't build with broadway or wayland since it wants the X-variant of
gdk. Fixes:
http://autobuild.buildroot.net/results/91d/91dc6ab749f967ae046f856f3f592d5a548499c6/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/tovid: Fix typo

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

package/minidlna: Fix typo

Fixes
http://autobuild.buildroot.net/results/5ac/5ac519320ec79cf8117060dcc2620aa517845000/

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

package/skeleton: also handle merged /usr in the staging

When the target uses a merged /usr setup, gdbserver will only report
paths in /lib to the remote gdb, which in turn will only look for
libraries in staging/lib and never in staging/usr/lib.

So. the merged (or non-merged) /usr setup must be replicated in the
staging.

The best solution where to do so is in the skeleton package, since it
is guaranteed to come before any package that installs things in the
staging, and even before the (internal or external) toolchain as well.

Reported-by: Pieterjan Camerlynck <pieterjan.camerlynck@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Pieterjan Camerlynck <pieterjan.camerlynck@gmail.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Peter Korsgaard <jacmet@uclibc.org>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/skeleton: parameterise the symlinks-or-mkdir macro

currently, we only make the merged symlinks or non-merged mkdirs in the
target directory. This is fine, as long as one does not run gdb.

However, as soon as one wants to run gdbserver on the target,
gdb will only search for libraries in /lib in the sysroot.

In preparation for a patch to fix that, make the symlink-or-mkdir macro
a function that takes as parameter the base directory where to create
symlinks or mkdir in.

Move the function declarations out of the custom/non-custom skeleton
conditional block, so they are always available.

Reported-by: Pieterjan Camerlynck <pieterjan.camerlynck@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Pieterjan Camerlynck <pieterjan.camerlynck@gmail.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Peter Korsgaard <jacmet@uclibc.org>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

opencv: use the same logic as opencv3 for gstreamer support

The logic for enabling gstreamer support in opencv is not correct. We
select gstreamer-0.10 packages but then we enable gstreamer1 support in
the opencv.mk file. opencv3 has the correct logic, so let's use it in
opencv as well.

Fixes:

  http://autobuild.buildroot.net/results/f6d/f6d0a4ffe347cbb868998856aca674ba8bc2e281/

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>