website/download: update daily snapshot section

This patch makes some minor changes to the download page of the website:
- don't capitalize Daily Snapshots as there is no reason to.
- remove extra link to daily snapshots, as this link is already available on
  the line below.
- add a link to the latest snapshot
- avoid 'Click here', especially if 'here' is not clickable

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

manual/getting started: rework 'System requirements' section

This patch reworks the requirements section of the manual as follows:
- some general rewording
- move configuration editor dependencies above the download tools, as this
  is the first thing people come in contact with.
- move sentence regarding -dev packages to configuration editor dependencies
  and restrict to 'libraries'.
- clarify the download tools part.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

manual/getting started: rework 'Community resources' section

This patch reworks the section 'Community resources' as follows:
- use a labeled list instead of subsections, as the content of each section
  is so small that it does not deserve a real section.
  Related to this, the anchors for each of these subsections are removed, as
  referencing such anchors that do not correspond to a section boundary are
  not displayed nicely.
- move the part about applying patches from patchwork to the Contributing
  section in the Developer Guide
- reword some parts of the text
- add link texts to some links, instead of showing the URL directly, to make
  the HTML / PDF display nicer.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

manual/getting started: rename and move 'Getting involved' section

The Getting Involved section is actually an overview of the ways to interact
with the Buildroot community, which is useful for developers _and_ users.
Therefore, this patch moves the section from the Developer Guide to the
Getting Started section, and renames it to 'Community Resources'.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

manual: remove 'Events' section

The list of buildroot events is outdated, a sign that the manual is not the
best place to refer to timely things like events. The wiki page is better
suited for such information.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

manual: high-level restructuring

The structure of the buildroot manual is not always clear. There is a large
number of chapters, and some chapters seem to overlap. The distinction
between general usage and developer information is not always clear.

This patch restructures the manual into four large parts:
- getting started
- user guide
- developer guide
- appendix

Except for the names of these parts, the section names are not yet changed.
Content-wise there are no changes yet either. This will be handled in
subsequent patches.

In order to achieve the introduction of a new level 'parts' above
'chapters', the section indicators (=, ==, ===, ...) of several sections
have to be moved one level down. Additionally, the leveloffset indication to
asciidoc has to be removed. Finally, to maintain more or less the same level
of detail in the table of contents, the toc.section.depth attribute is
reduced as well. Note that for some sections, less detail is visible now.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

dependencies: Reject gcj in BR2_NEEDS_HOST_JAVA check

Xbmc uses the Groovy java package to generate API code, for details see
http://wiki.xbmc.org/index.php?title=Codegeneration#How_it_works

Groovy itself is broken when used with gcj:
http://jira.codehaus.org/browse/GROOVY-3701

Xbmc recommends to use another java package instead, like openjdk:
http://forum.xbmc.org/showthread.php?tid=199742

Until we come up with a host package to provide a java setup and
in order to make autobuilder happy let´s reject gcj when we check
for a java installation on the host.

Currently Xbmc is the only package using BR2_NEEDS_HOST_JAVA
so changing the behaviour of dependencies.sh has no side effects.

Fixes
http://autobuild.buildroot.net/results/213/213d7454a1dc68cfb8f4c8d9508f52ebf8e58e33/
http://autobuild.buildroot.net/results/639/639cda872cf36c122cbd6139dd76f6b51ac1864e/
http://autobuild.buildroot.net/results/c9e/c9eb12c95b2939c64590692a82af8d84bbd90956/

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

package/nut: fix parallel install

Fixes:
    http://autobuild.buildroot.net/results/256/2567e13cd5bc702bc3a38a1d6fc8e34022cc7db5/

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

package/oprofile: fix configure test with gcc 4.9.x

Backport the upstream patch that fix a configure test breaked with gcc 4.9.x.

This patch must be removed with the next version bump.

Fixes:
http://autobuild.buildroot.net/results/f80/f8083f038d0efb7b5becf7810b6fcaf440fb8294/build-end.log

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

CHANGES: update with recent changes

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

gcc/4.9: fix C++ exceptions and pthread_exit()

Following the introduction of the support for the musl C library, the
support of C++ exceptions or features like pthread_exit() got broken
even with other libraries such as glibc. This was reported as bug #7028.

The problem was caused by the gcc patch needed to add support for
musl, which modified the libgcc/unwind-dw2-fde-dip.c logic to decide
whether USE_PT_GNU_EH_FRAME should be enabled or not. It completely
removed the existing logic, replacing it by a single logic based on
the definition of TARGET_DL_ITERATE_PHDR. However, this constant gets
defined by the configure script only for Solaris, or Linux Musl
platforms. For glibc/uClibc, the configure script does not define it,
and therefore USE_PT_GNU_EH_FRAME is not defined, causing issues with
exception handling.

This patch fixes that by restoring all the logic of
libgcc/unwind-dw2-fde-dip.c, and just adding the musl logic as one
more case.

It has been successfully runtime tested using the two code examples
provided in bug #7208, with uClibc, musl and glibc.

Cc: Krzysztof Wrzalik <kwrzalik@gmail.com>
Cc: David Bachelart <david.bachelart@bbright.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcc/4.8: fix C++ exceptions and pthread_exit()

Following the introduction of the support for the musl C library, the
support of C++ exceptions or features like pthread_exit() got broken
even with other libraries such as glibc. This was reported as bug #7028.

The problem was caused by the gcc patch needed to add support for
musl, which modified the libgcc/unwind-dw2-fde-dip.c logic to decide
whether USE_PT_GNU_EH_FRAME should be enabled or not. It completely
removed the existing logic, replacing it by a single logic based on
the definition of TARGET_DL_ITERATE_PHDR. However, this constant gets
defined by the configure script only for Solaris, or Linux Musl
platforms. For glibc/uClibc, the configure script does not define it,
and therefore USE_PT_GNU_EH_FRAME is not defined, causing issues with
exception handling.

This patch fixes that by restoring all the logic of
libgcc/unwind-dw2-fde-dip.c, and just adding the musl logic as one
more case.

It has been successfully runtime tested using the two code examples
provided in bug #7208, with uClibc, musl and glibc.

Cc: Krzysztof Wrzalik <kwrzalik@gmail.com>
Cc: David Bachelart <david.bachelart@bbright.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcc/4.7: fix C++ exceptions and pthread_exit()

Following the introduction of the support for the musl C library, the
support of C++ exceptions or features like pthread_exit() got broken
even with other libraries such as glibc. This was reported as bug #7028.

The problem was caused by the gcc patch needed to add support for
musl, which modified the libgcc/unwind-dw2-fde-dip.c logic to decide
whether USE_PT_GNU_EH_FRAME should be enabled or not. It completely
removed the existing logic, replacing it by a single logic based on
the definition of TARGET_DL_ITERATE_PHDR. However, this constant gets
defined by the configure script only for Solaris, or Linux Musl
platforms. For glibc/uClibc, the configure script does not define it,
and therefore USE_PT_GNU_EH_FRAME is not defined, causing issues with
exception handling.

This patch fixes that by restoring all the logic of
libgcc/unwind-dw2-fde-dip.c, and just adding the musl logic as one
more case.

It has been successfully runtime tested using the two code examples
provided in bug #7208, with uClibc, musl and glibc.

Cc: Krzysztof Wrzalik <kwrzalik@gmail.com>
Cc: David Bachelart <david.bachelart@bbright.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcc/arc-4.8-R3: add patch to enable more C++ features with uClibc

This commit fixes bug #7250, by allowing more libstdc++ features to be
enabled with uClibc. libstdc++ wants an absolutely complete C99
support in the C library before enabling *any* feature that needs some
C99 functions. However, uClibc doesn't provide C99 complex numbers, so
libstdc++ disables a lot of C++ standard methods, even though they are
not related to C99 complex numbers.

A partial solution already existed in the patch
302-c99-snprintf.patch, but this commit replaces it by the more
complete 850-libstdcxx-uclibc-c99.patch, which is highly inspired by
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58393, except that it
doesn't rely on configure.ac checks, but simply on testing
defined(__UCLIBC__) like was done in 302-c99-snprintf.patch. This
allows to avoid having to autoreconf gcc, which is quite complicated
to achieve.

Reported-by: Richard <tarka.t.otter@gmail.com>
Cc: Richard <tarka.t.otter@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcc/4.9: add patch to enable more C++ features with uClibc

This commit fixes bug #7250, by allowing more libstdc++ features to be
enabled with uClibc. libstdc++ wants an absolutely complete C99
support in the C library before enabling *any* feature that needs some
C99 functions. However, uClibc doesn't provide C99 complex numbers, so
libstdc++ disables a lot of C++ standard methods, even though they are
not related to C99 complex numbers.

A partial solution already existed in the patch
302-c99-snprintf.patch, but this commit replaces it by the more
complete 850-libstdcxx-uclibc-c99.patch, which is highly inspired by
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58393, except that it
doesn't rely on configure.ac checks, but simply on testing
defined(__UCLIBC__) like was done in 302-c99-snprintf.patch. This
allows to avoid having to autoreconf gcc, which is quite complicated
to achieve.

Reported-by: Richard <tarka.t.otter@gmail.com>
Cc: Richard <tarka.t.otter@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcc/4.8: add patch to enable more C++ features with uClibc

This commit fixes bug #7250, by allowing more libstdc++ features to be
enabled with uClibc. libstdc++ wants an absolutely complete C99
support in the C library before enabling *any* feature that needs some
C99 functions. However, uClibc doesn't provide C99 complex numbers, so
libstdc++ disables a lot of C++ standard methods, even though they are
not related to C99 complex numbers.

A partial solution already existed in the patch
302-c99-snprintf.patch, but this commit replaces it by the more
complete 850-libstdcxx-uclibc-c99.patch, which is highly inspired by
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58393, except that it
doesn't rely on configure.ac checks, but simply on testing
defined(__UCLIBC__) like was done in 302-c99-snprintf.patch. This
allows to avoid having to autoreconf gcc, which is quite complicated
to achieve.

Reported-by: Richard <tarka.t.otter@gmail.com>
Cc: Richard <tarka.t.otter@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcc/4.7: add patch to enable more C++ features with uClibc

This commit fixes bug #7250, by allowing more libstdc++ features to be
enabled with uClibc. libstdc++ wants an absolutely complete C99
support in the C library before enabling *any* feature that needs some
C99 functions. However, uClibc doesn't provide C99 complex numbers, so
libstdc++ disables a lot of C++ standard methods, even though they are
not related to C99 complex numbers.

A partial solution already existed in the patch
302-c99-snprintf.patch, but this commit replaces it by the more
complete 850-libstdcxx-uclibc-c99.patch, which is highly inspired by
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58393, except that it
doesn't rely on configure.ac checks, but simply on testing
defined(__UCLIBC__) like was done in 302-c99-snprintf.patch. This
allows to avoid having to autoreconf gcc, which is quite complicated
to achieve.

Reported-by: Richard <tarka.t.otter@gmail.com>
Cc: Richard <tarka.t.otter@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/{atngw100, atstk100x}: fix build of AVR32 defconfigs

Since we marked the AVR32 architecture deprecated, it can no longer be
selected in a normal configuration. This has caused the AVR32
defconfigs to fail to build in our daily Jenkins tests. In order to
make them build again, they must use BR2_DEPRECATED=y. Of course, when
we remove AVR32 completely, we'll remove those defconfigs as well.

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

libnftnl: add patch to avoid symbol conflict

When ELF binaries and shared libraries are used, the internal
functions of libnftnl such as xfree() are not visible to the outside
world (their visibility is 'hidden'). Therefore, the fact that other
programs (especially nftables) may have symbols with the same name
does not cause any problem.

However, when doing static linking on a non-ELF platform (such as
Blackfin, which uses the FLAT binary format), there is no way of
encoding this visibility. Therefore, the xfree() symbols of libnftnl
becomes visible to the outside world, causing a conflict with the
xfree() symbol defined by nftables.

To solve this, this patch renames the libnftnl xfree() function to
libnftnl_xfree().

This commit fixes a problem seen after building nftables statically on
Blackfin. This problem was seen after the previous patch fixing the
autobuilder failure:

  http://autobuild.buildroot.org/results/98b/98b707ffdeeb1cda94b7c1019ef29cf5fd7db8bf/

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

nftables: fix build in static scenarios

When linking against readline, it forgets to link against ncurses,
which is needed by readline. Fix this by passing LIBS="-lcurses" to
the configure script.

Fixes:

  http://autobuild.buildroot.org/results/98b/98b707ffdeeb1cda94b7c1019ef29cf5fd7db8bf/

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

package/empty: Respect LDFLAGS, fix link order

Fixes
http://autobuild.buildroot.net/results/50f/50fedb54cf005d1177d85c5a34fd10a1c354ba5b/
http://autobuild.buildroot.net/results/9fc/9fce0825772e99de3c6808b729641147385413aa/

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

qmx6/wandboard_defconfig: switch to custom kernel headers version

Since 3.0.x was EOL and deprecated this would cause it to use newer
headers than the kernel and cause issues.
So switch to custom kernel headers version string to fix this.

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

linux-headers: mark EOL versions deprecated

Follow the upstream EOL cycle and mark 3.0, 3.11, 3.13 and 3.15
as deprecated for 2014.08.

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

manual-text: hide image representations

Images specified in the asciidoc sources (currently only the logo) are
displayed as the file name in the text version of the manual. This causes an
odd line to appear:
    logo.png

Avoid this by setting the image representation macros to {empty} in
manual-text.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Samuel Martin <s.martin49@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

manual-text: make sure URLs are displayed if a link text was provided

When the asciidoc source contain URLs of the form:
    http://example.com[An example website]
the text representation of the manual would only contain:
    An example website
without displaying the actual URL.

This patch adds an asciidoc configuration file that sets the inline macros
for several URL types so that the display becomes:
    An example website [http://example.com]
For URLs where no link text was provided, the display becomes:
    http://example.com
which is the same as before.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Samuel Martin <s.martin49@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gendoc infra: add support for asciidoc configuration files

This patch introduces support for asciidoc configuration files, specific for
each output format (html, text, pdf, ...). This is useful to make certain
tweaks in the document generation.

If a file docs/manual/asciidoc-<format>.conf is present, it is passed to
asciidoc as configuration file. If no file for the current format is
present, the options passed to asciidoc are empty.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Samuel Martin <s.martin49@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gendoc infra: use $(pkgname) instead of explicitly passing 'manual'

In the gendoc infrastructure, using an assignment of the form
    FOO = docs/$(1)/bar
inside GENDOC_INNER does not work as expected: the $(1) value is empty here
and the value of FOO becomes 'docs//bar'.

Parameters $(2), $(3), etc. do not have this problem. The specific thing
about $(1) is that it is a parameter to GENDOC itself (indicating the
document to create) and passed transparently to GENDOC_INNER.

This is different from the package infrastructures, where $(1) is set from
$(pkgname). In fact, the same strategy could be used by the gendoc
infrastructure as well, as $(pkgname) resolves to 'manual' for file
docs/manual/manual.mk. This has the advantage that the described problem
does not occur.

Note that this means that if we want to use the same GENDOC infrastructure
for another document, it will have to reside in a separate directory than
the manual.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Makefile: unconditionally include pkg-utils.mk

Currently, pkg-utils.mk (included via package/Makefile.in) is only included
when a configuration file already exists. This means that none of the
utilities it defines are available without .config.

In particular:
- the MESSAGE macro, causing pretty build output. Since some make targets
  can be run even without .config, like 'make manual', not having this
  pretty printing is odd.

- pkgname, pkgdir: in a subsequent patch, these functions will be used for
  the generation of the manual, and since this should work also without
  .config, we need these functions to be available.

This patch moves the include of pkg-utils.mk from package/Makefile.in to
Makefile, outside of the check for .config.

This is a quick fix. The full solution involves to minimize the amount of
Makefile code that is guarded by a check on .config. This approach will be
taken in the 2014.11 release cycle.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

perl-net-ssleay: fix build

the new option OPTIMIZE needs the same replacement.

see http://autobuild.buildroot.net/results/278f9d121273a16cedc009c0b8df0a8659d8fb73/

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

package/iftop: fix static build

Use pcap-config to list optional libpcap dependencies that we need to list
when building statically.

Inspired by Baruch Siach
http://git.buildroot.net/buildroot/commit/package/dhcpdump/dhcpdump.mk?id=429f4415cd153c6809394a8b3245d4d15bba3ec3

Fixes
http://autobuild.buildroot.net/results/c7e/c7e3b2897a9cb9ab55dc7b1a2cd1961235d2d1a4/
http://autobuild.buildroot.net/results/60f/60f82bedae255f6b69c9a5ac22686c76c6276301/
http://autobuild.buildroot.net/results/1b7/1b771af04a95a78144141a5d555c97bbb5a7e13f/
http://autobuild.buildroot.net/results/1f4/1f45e2adcc80c4a209aa5895260985460933575c/

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

package/ngrep: fix static build

Use pcap-config to list optional libpcap dependencies that we need to list
when building statically.

Inspired by Baruch Siach
http://git.buildroot.net/buildroot/commit/package/dhcpdump/dhcpdump.mk?id=429f4415cd153c6809394a8b3245d4d15bba3ec3

Fixes
http://autobuild.buildroot.net/results/3c5/3c584b850cabebcf93dfd61c59e28988165a41c4/
http://autobuild.buildroot.net/results/938/93842b54767cca51c68ad33ddc93ec58d70602c8/
http://autobuild.buildroot.net/results/0ae/0ae4adec7fe66f819c7b3dabf13648fb29247de6/
http://autobuild.buildroot.net/results/9c2/9c203bc6ca91da68c9e7335611c1fb56224380bc/

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

xapp_rstart: Depends on MMU

Fixes
  CCLD   rstartd.real
auth.o: In function `_do_auth':
auth.c:(.text+0x6a): undefined reference to `_fork'
server.o: In function `_detach':
server.c:(.text+0x80a): undefined reference to `_fork'

To reproduce use this defconfig
http://autobuild.buildroot.net/results/8d5/8d5ad404fab7ec2a501972ba9c08b0109e420b80/defconfig

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

package/matchbox: Depends on MMU

Fixes matchbox-panel:
http://autobuild.buildroot.net/results/8d5/8d5ad404fab7ec2a501972ba9c08b0109e420b80/

and matchbox-desktop:
mbdesktop_win_plugin.o: In function `mbdesktop_win_plugin_load':
/home/br/br/output/build/matchbox-desktop-0.9.1/src/mbdesktop_win_plugin.c:22: undefined reference to `_fork'

and matchbox-wm:
matchbox-remote.o: In function `mbcommand':
/home/fli4l/br3/output/build/matchbox-wm-1.2/src/matchbox-remote.c:147: undefined reference to `_fork'

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

package/webkit: select cairo, since we select one of its options

When selecting an option of a package, that package must also be
selected, otherwise the package is not built.

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

package/infozip: Rename HOST_CC to HOSTCC

Wrong variable name is used which can lead to build errors for the host variant:

[...] -I. -DUNIX" AS=" -c" -f unix/Makefile generic

c _match.s
make[1]: c: command not found

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

systemd: disable link time optimization

As reported in https://bugs.busybox.net/show_bug.cgi?id=7286,
systemd fails to build with BR2_OPTIMIZE_2=y. This mode activates the GCC
options -flto and -ffat-lto-objects, which do not mix well with the GCC
tricks used in src/compat-libs/linkwarning.h to build the compatibility
libraries.

As a temporary workaround, this patch disables link time optimization for
systemd.

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gnupg2: Needs a suitable iconv implementation

gnupg2 needs a toolchain with locale support or a package which provides
a suitable iconv implementation (libiconv). Otherwise it will fail at
the configure phase with an error like this one:

*** It is now required to build with support for iconv
*** Please install a suitable iconv implementation.

Fixes:
  http://autobuild.buildroot.net/results/8c9/8c93c28533dfebffa8b2e34b1421d3fa3cdeb278/

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

subversion: security bump to version 1.7.18

Fixes:

CVE-2014-0032 - mod_dav_svn is vunerable to a remotely triggerable
segfault DoS vulnerability when SVNListParentPath is on.

CVE-2014-3522 - Serf RA layer does not correctly validate certificates
with wildcards in them for HTTPS.

CVE-2014-3528 - Credentials cached with Subversion may be sent to the
wrong server.

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

gd: add patch for CVE-2014-2497

Fixes CVE-2014-2497 - NULL pointer dereference
Patch from upstream:
https://bitbucket.org/libgd/gd-libgd/commits/463c3bd09bfe8e924e19acad7a2a6af16953a704

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

package/localedef: don't use config.site when cross-compiling

On fedora 20 64bits host, the file /usr/share/config.site contains
a fix for installing libraries into /lib/lib64 on 64bits systems
that redefine libdir in the generated Makefile.

This causes the bug 7262 reported Normen Bolling.

Set CONFIG_SITE to "no" to not use config.site

fixes: https://bugs.busybox.net/show_bug.cgi?id=7262

Reported-by: Normen Bolling <normen.bolling@teufel.de>
Cc: Normen Bolling <normen.bolling@teufel.de>
Cc: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Tested-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/tar: Depends on MMU

Fixes
misc.o: In function `_xfork':
misc.c:(.text+0x614): undefined reference to `_fork'
../lib/libtar.a(rtapelib.o): In function `_rmt_open__':
rtapelib.c:(.text+0x6d2): undefined reference to `_fork'

To reproduce use this defconfig
http://autobuild.buildroot.net/results/8d5/8d5ad404fab7ec2a501972ba9c08b0109e420b80/defconfig

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

gst1-plugins-good: v4l2 plugin needs kernel headers 3.0+

Fixes http://autobuild.buildroot.net/results/284/284313e660c3c9fad891fd71aa56ca5751946dc6/

For the multi planar api (mp_pix). It was actually added in 2.6.39 (commit
f8f3914cf9), but we only have symbols for 3.x kernels, so use 3.0 instead.

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

qt: don't install translation files by default

Commit 93917b6980f7f2b51302e1a3fa451b07cf7d674e (2013.11) introduced
the installation of the binary .qm translation files, unconditionally.

However, the installed size is very large for some embedded systems
(about 8 MB), and they are not needed in many cases.

In order to avoid such a waste of storage space, commit
2ff329412f490dc916bab3377073ef5950a46298 (2014.08) made this
installation an option that could be disabled. For backward
compatibility, the option default was set to yes.

Given the size of these files, and that they had never been installed
by Buildroot versions before 2013.11, change the default to not
installing them.

Interested users can still activate it as needed.

Also update and improve the help text.

Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Reviewed-by: Danomi Manchego <danomimanchego123@gmail.com>
Cc: Danomi Manchego <danomimanchego123@gmail.com>
Cc: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libcuefile: fix static linking

Add a patch to libcuefile to fix BR2_PREFER_STATIC_LIB=y configurations.

Fixes:

  http://autobuild.buildroot.org/results/13e/13ecdab5bb132532674a51e0e0e1136d417dfcd9/

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

libebml: properly support static only builds

This commit adjusts the libebml package to not build a shared library
when BR2_PREFER_STATIC_LIB=y.

Fixes:

  http://autobuild.buildroot.org/results/262/26279f72b6c5708f363b3cce2102afa0eccc3a72/

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

bandwidthd: incorporate fixes for autobuilders

"The build failures of bandwidthd observed in the autobuilders in
relation to libpng are caused by the usage of -L/usr/lib
-L/usr/local/lib in the compiler flags, which leads the configure
script to try to link against a host libpng library."

These issues are fixed by:

* Version bump to -r08
"[patched] the configure.ac script to remove the hardcoded
-L/usr/local/lib and -I/usr/local/include. The bandwidthd package
-was already autoreconfigured, so there is no need to add it."

* "Adding the --without-x option, which ensures that the configure.ac
script will not add -L/usr/lib to the compiler flags (and we anyway
haven't added any sort of X.org support to bandwidthd for the
moment)."
Both fixes by Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

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

package/ffmpeg: prevent stripping the ffmpeg binary

Packages shouldn't strip executables by themselves, so that Buildroot
controls whether stripping should occur or not. This also fixes the
build on Blackfin FLAT where stripping actually doesn't work because
stripping FLAT binaries is not supported.

Inspired by Thomas:
http://git.buildroot.net/buildroot/commit/package/hdparm/hdparm.mk?id=11dbb615338cd768acc2c3138064ddf557e9b60b

Fixes
http://autobuild.buildroot.net/results/581/581c43b50c9717e53334eb156ece95fe499ccb04/
http://autobuild.buildroot.net/results/57a/57a1bdc22b803114c018fa007884385bf8355151/
http://autobuild.buildroot.net/results/d8a/d8a1aadebd302abc923bc4865e1242ab2479d641/

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

pkg-perl: fix build of perl XS packages

Set OPTIMIZE in order to avoid the use of the host $Config{optimize} value
which could cause breakage when cross-compiling for the target.

See discussion in http://lists.busybox.net/pipermail/buildroot/2014-August/104129.html

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

bluez5_utils: gatttool depends on bluez5_utils client

The gatttool utility uses readline, and is therefore built only if
--enable-client is used, i.e when bluez5_utils client applications are
enabled. Therefore, this commit makes
BR2_PACKAGE_BLUEZ5_UTILS_GATTTOOL depend on
BR2_PACKAGE_BLUEZ5_UTILS_CLIENT. It also moves
BR2_PACKAGE_BLUEZ5_UTILS_GATTTOOL to be right below the
BR2_PACKAGE_BLUEZ5_UTILS_CLIENT definition.

Fixes:

  http://autobuild.buildroot.org/results/f2b/f2b8b1bf5d50e6223a9a6d6858320975edb2a25a/

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

linux: bump default to version 3.16.1

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

linux-headers: bump 3.{4, 10, 14, 15, 16}.x series

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

ltrace: disable on mips/mipsel

ltrace currently doesn't build on mips/mipsel, and it's an upstream
issue that has been reported. Until it get fixed, let's disable ltrace
for mips/mipsel.

Fixes:

  http://autobuild.buildroot.org/results/43a/43a8fc7075f52eab74ebfee4c9f25dd2b886e75e/

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

synergy: needs mmu

Fixes:
http://autobuild.buildroot.net/results/8d4/8d435c1605204f154c8a0a94ebc06d9c29a1768a/

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

pppd: security bump to version 2.4.7

Fixes CVE-2014-3158 - potential security issue in parsing option files.

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

tftpd: depends on MMU

This package needs MMU. Hence added dependency on BR2_USE_MMU

Fixes:
http://autobuild.buildroot.net/results/3e1/3e16d8c9ae3b77393e082c3dd9b7b3de9aa040c0//

Signed-off-by: Yuvaraj Patil <yuvaraj.patil@wipro.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

procps-ng: Depends on MMU

This package needs MMU. Hence added dependency on BR2_USE_MMU

Fixes:
http://autobuild.buildroot.net/results/f24/f24aa16857e91fb6b03f9ac7ab0442fedcab8546//

Signed-off-by: Yuvaraj Patil <yuvaraj.patil@wipro.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/xbmc: Bump version to 13.2rc1

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

fbv: fix typo in pkg-config variable

Reported-by: Danomi Manchego <danomimanchego123@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/fbv: Fix static linking against libpng

Fixes
http://autobuild.buildroot.net/results/9ca/9cacec257f979745a7d4a43224b243ee4e16f9ac/
http://autobuild.buildroot.net/results/996/996e42396546555e9fcf0c7135aac610ccca4eee/
http://autobuild.buildroot.net/results/c91/c918046900045c34c73e16578026cce3e6bcccf8/

[Peter: adjust description, pass --static to pkg-config, only use for static linking]
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/fbv: Typo fix

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

perl: fix warnings at runtime

autosplit.ix files are used by AutoLoader.pm

    Can't locate auto/Net/SSLeay/autosplit.ix in @INC (@INC contains: /usr/lib/perl5/site_perl/5.18.2/arm-linux /usr/lib/perl5/site_perl/5.18.2 /usr/lib/perl5/5.18.2/arm-linux /usr/lib/perl5/5.18.2 .) at /usr/lib/perl5/5.18.2/AutoLoader.pm line 181.
     at /usr/lib/perl5/site_perl/5.18.2/arm-linux/Net/SSLeay.pm line 24.

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

postgresql: disable spinlock support for nios2

Fixes:
http://autobuild.buildroot.net/results/4d2/4d2f962ae59cfe11fce4d6790dbccfbb527a8f5f/

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

espeak : depends on MMU

Fixes http://autobuild.buildroot.net/results/dae/dae5c59dfa127e1c3f9e3885cdd1b97e54d8fe41/
and http://autobuild.buildroot.net/results/fa0/fa04a0fe7d7a201c5f733bd798c0da50ffb81271/
espeak uses fork(), don't work on non-MMU platforms

[Peter: indent with tab instead of spaces]
Signed-off-by: Arnaud Aujon <arnaud@intelibre.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

network-manager: fix build on toolchains not exporting CLOCK_BOOTTIME (E.G. uClibc)

Fixes
http://autobuild.buildroot.net/results/c13/c1388cf73961ddf06f771afe77a7e1e9fc0d8c20/
http://autobuild.buildroot.net/results/d33/d33ed1c3bf1be7c8f8ea02590a770b0d29c7d9fa/
http://autobuild.buildroot.net/results/093/093656db99067eb7bff725e292f8fc880a1d748e/

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

openssl: add fix for hash cryptodev offloading

See http://rt.openssl.org/Ticket/Display.html?id=2770&user=guest&pass=guest
This has been sitting for ages in the openssl tracker and it's verified
to cause issues.
The patch only touches cryptodev engine offloading so it's pretty safe.
Tested on CAAM SEC4 hardware.

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

Update for 2014.08-rc1

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

CHANGES: add highlights for 2014.08 release

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

linux-headers: bump 3.{4, 10, 14, 15}.x series

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

dropbear: bump to version 2014.65

Fixes some regressions introduced by 2014.64

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

wireshark: security bump to version 1.10.9

Fixes:
WNPA-SEC-2014-08 / CVE-2014-5161 / CVE-2014-5162
The Catapult DCT2000 and IrDA dissectors could underrun a buffer.

WNPA-SEC-2014-09 / CVE-2014-5163
The GSM Management dissector could crash.

WNPA-SEC-2014-10 / CVE-2014-5164
The RLC dissector could crash.

WNPA-SEC-2014-11 / CVE-2014-5165
The ASN.1 BER dissector could crash.

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

CHANGES: update with recent changes

The description of important changes still needs to be added.

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

dhcpdump: Fix strsep() feature test

Use the official _BSD_SOURCE feature test macro instead of the meaningless
HAVE_STRSEP macro in order to detect the availability of strsep().

This allows toolchains supporting strsep() to use it instead of the custom
implementation from dhcpdump, which also avoids the following error with some
toolchains:

In file included from dhcpdump.c:30:0:
dhcpdump.c: At top level:
strsep.c:65:23: error: register name not specified for ‘delim’
  register const char *delim;
                       ^

Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

dhcpdump: Fix patch file name

Use the naming convention for patches, which is:

<packagename>-<number>-<description>.patch

Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

qemu/aarch64-virt: add new sample config

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

linux: fix LINUX_SITE for release candidates

Signed-off-by: Jean-Baptiste Theou <jtheou@adeneo-embedded.us>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

linux-headers: bump 3.2.x series

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

openssl: security bump to version 1.0.1i

Fixes:
CVE-2014-3508 - Information leak in pretty printing functions
CVE-2014-5139 - Crash with SRP ciphersuite in Server Hello message
CVE-2014-3509 - Race condition in ssl_parse_serverhello_tlsext
CVE-2014-3505 - Double Free when processing DTLS packets
CVE-2014-3506 - DTLS memory exhaustion
CVE-2014-3507 - DTLS memory leak from zero-length fragments
CVE-2014-3510 - OpenSSL DTLS anonymous EC(DH) denial of service
CVE-2014-3511 - OpenSSL TLS protocol downgrade attack
CVE-2014-3512 - SRP buffer overrun

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Tested-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

mongoose: fix build failure

mongoose.h looks missing because we're not building locally, so add
the appropriate -I flag. Fixes:
http://autobuild.buildroot.net/results/701/701711626548bd166cd5bc5669e4761ffed074d1/

[Thomas: change the solution to use -I$(@D) instead of switching to
the build directory.]

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

toolchain: fix 3.16 selection for custom version

Fix d'oh typo from 911516f1b6dec4b2985a1f454949f8d9d10f6ef5 that
prevents custom version string 3.16 headers from working properly.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/imx233_olinuxino: bump to kernel 3.16

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

package/xbmc: Bump version to 13.2b3

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

configs/qemu-microblaze: lock down kernel headers

Now that we use an internal toolchain as default care should be taken to
avoid breaking things.

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

mongoose: patch to avoid _LARGEFILE_SOURCE redefinition warning

Signed-off-by: Davide Viti <zinosat@tiscali.it>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

mongoose: bump to version 5.4

Avoid compilation of the server via the upstream Makefile which
unconditionally uses -lssl

The -lssl flag is appended to MONGOOSE_CFLAGS whenever
BR2_PACKAGE_OPENSSL=y gets set.

Signed-off-by: Davide Viti <zinosat@tiscali.it>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python: renumber patch

Advance the serial number of a patch from 015 to 016 to avoid collision and
confusion.

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

python3: fix termios build for xtensa

Forward port the patch from c821210b1bfa (python: fix termios build for
xtensa) to python3.

Fixes:
http://autobuild.buildroot.net/results/e7f/e7f71fbcb0796f315052947382fa39d85c5dc8fe/

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

nodejs: updated to 0.10.30

Signed-off-by: Jordi Llonch <jordi.llonch@rochsystems.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs: bump kernel used for the RPi

The rpi-3.12.y branch has now been merged with the 3.12.26 upstream.

Bump kernel headers used for the toolchain at the same time.

Fixes
http://jenkins.free-electrons.com/job/buildroot/config=raspberrypi_defconfig/320/console

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

infra/pkg-kconfig: incorporate oldconfig call to avoid endless rebuilds

The configure step of the busybox package performs 'make oldconfig', which
causes the .config file to be updated. Thus, the .config file is more recent
than our stamp file .stamp_kconfig_fixup_done. On a subsequent build, our
dependency rules would kick in, and run the config fixup again, thus
kicking in the package's configure, build and install steps yet once
more, that, ad infinitum.

One solution is to modify kconfig-package to introduce an explicit touch of
the .kconfig_fixup_config_done stamp file, as post-configure hook.

Another solution, implemented by this patch, is to move the oldconfig call
from the package's .mk file to the kconfig-package infrastructure and make
sure it is done as part of the fixup commands. This way, the stamp file will
only be touched once, after the full fixup (including oldconfig) and no
endless rebuilds will occur.

Reported-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Suggested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

infra/pkg-kconfig: join comments to the related code-block

The current layout of comments wrt. to the related code-blocks is a bit
hard to handle visually.

Remove empty lines between the comments and their respective related
code-blocks.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

busybox: enable ash getopts

It's increasingly used by scripts like some in the btrfs-progs package.

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

gcc 4.9.1: add patch for PR60102

Fixes ICE for SPE ABI, see:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60102

And re-enable gcc 4.9.1 for SPE.

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

manual: fix copy/paste error in python package example

The SOURCE references "LIBFOO_VERSION", should be "PYTHON_FOO_VERSION".

Signed-off-by: Danomi Manchego <danomimanchego123@gmail.com>
Acked-by: Thomas De Schampheleire < thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

espeak: bump to version 1.48.04

* Bump to latest upstream version 1.48.04
* Remove patch as applied upstream

Signed-off-by: Arnaud Aujon <arnaud@intelibre.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

espeak: depends on threads

Fixes autobuild failure : http://autobuild.buildroot.org/results/f39/f3903d2b4a16226df2d460dcd2e26c7fb87613ab

Signed-off-by: Arnaud Aujon <arnaud@intelibre.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gdb: allow building on AArch64 systems

gdb has had AArch64 support since 7.6 which is now the default minimum
version of GDB in the tree. Older versions are now legacy.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

iproute2: bump to version 3.16.0

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Tested-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

rpi-userland: .mk: Indent rules with tabs

Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

rpi-userland: Bump version

Bump version to f9e6f9f3def8dc18dc0092cd695ccf53b8ba3efe and revert local commit
f22830e456833e98e11f09d2aa38017f5aaa557c 'rpi-userland: add patch fix missing
libvcilcs install rules' in order to replace the local patch from this commit by
the upstream commit 1ecc49c97b7be3d524cb4cfc97401f16b485a4ee 'vmcs_host: Fix
missing install of libvcilcs.so'.

Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

jack2: add staging installation

jack2 installs some headers and libraries that can be used by other
applications, so it makes sense to install it to the staging
directory.

Signed-off-by: Pieter De Gendt <pieter.degendt@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>