To allow proper nesting of suboptions of a package, the suboptions should
come directly after the main option, and cannot be interleaved with a
comment.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
alsa-utils uses fork which is not available on targets without MMU support.
It seems to be possible to replace fork with vfork in alsa-utils, but we do
not like to carry such patches in buildroot without them being accepted
upstream.
Fixes
http://autobuild.buildroot.org/results/9f8/9f8e572c9f1c677155cc7d1828371bcf572ff878
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The AM_PATH_ALSA macro in utils/alsa.m4 unconditionally uses -ldl. This
breaks compilation of alsa-utils (and probably other packages using this
macro) for targets that do not support dynamic loading, such as for
Blackfin FLAT binaries.
This patch updates the macro to check if dlopen is available, and use that
result to conditionally add -ldl to the list of libraries.
Fixes
http://autobuild.buildroot.org/results/de2/de286880973be956f6c504aa9a758171d6f49674/build-end.log
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The FLAT GNU toolchain doesn't include the dlfcn.h header file.
Provide the necessary declarations (RTLD_*) to make alsa-lib happy.
Fixes
http://autobuild.buildroot.org/results/706/7069e1f43cbed745d65f7dd9904a3fff034530ac/build-end.log
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
[Thomas: change sequence number from 003 to 0003, update patch and commit
message ]
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
In commit a9baea4345 ('pixman: add patch
to fix Microblaze build failure'), a patch is added to the pixman
package to avoid using the FE_DIVBYZERO definition when it is not
available. However, it was using the have_fe_divbyzero variable to
define or not HAVE_FEDIVBYZERO, while the AC_CHECK_DECL autoconf macro
sets the ac_cv_have_decl_FE_DIVBYZERO variable.
The end result was that the FE_DIVBYZERO macro was considered as never
being available. This commit fixes that by using the appropriate
variable.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Some packages (actually, just oprofile) need to link against libiberty.
This option just installs libiberty.a so it has no effect on the target,
therefore it's not needed to add a config option for it.
Before binutils-2.24, there was a bug in libiberty/Makefile.in that
caused libiberty to be installed regardless of the
--enable-install-libiberty option. This problem wasn't noticed before
because binutils-2.24 is not selected on any of the autobuilders: the
version can only be selected if an internal toolchain is used, and it
defaults to 2.21.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The Microblaze build of pixman was failing due to FE_DIVBYZERO not
being implemented. It turns out that the usage of it, like fenv.h and
feenableexcept() is optional. So the patch simply adds a configure
check and disables the appropriate code (which is only use in the
tests anyway).
This commit also renames the existing patch to follow the patch naming
convention, and get a reliable ordering when applying patches.
Fixes:
http://autobuild.buildroot.org/results/806/8064092cdbac85fbf4322429d29d5d11dc51860f/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As our architecture support expands to a number of architectures that
do not implement NPTL threading, and the number of packages that
depend on NPTL specific features, it has become necessary to be able
to know whether the toolchain has NPTL support or not.
This commit adds a new BR2_TOOLCHAIN_HAS_THREADS_NPTL hidden Config.in
option that allows packages to know whether NPTL is available or not.
This hidden option is:
* Automatically enabled when glibc/eglibc or musl toolchains are
used, either internal or external.
* Automatically enabled when an internal uClibc toolchain with NPTL
support is configured. It is left disabled otherwise for internal
uClibc toolchains.
* Configured according to a visible Config.in option for custom
external uClibc toolchains.
[Peter: factor _EXTERNAL_HAS_THREADS in single if as suggested by Arnout]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
libcgi always builds both the shared and static library, which doesn't
work on architecture that don't support shared library at all, such as
Blackfin with the FLAT format. libcgi uses autoconf, but not automake,
and the Makefile is not of the highest possible quality, so this
commit fixes the problem by introducing a "STATIC" variable that can
be set from the environment. When set to a non-empty value, the
Makefile assumes it should only build the static version of the
library.
Note that this package is in desperate need of some care: there is one
single patch that mixes several changes together, it doesn't have a
description or a Signed-off-by line, and there is now a github project
for libcgi at https://github.com/rafaelsteil/libcgi/ that has the same
fixes.
However, for the purpose of the master branch, we're doing the most
minimal fix possible, by just adding this STATIC variable.
Fixes:
http://autobuild.buildroot.net/results/625/625105bcaf26345f422b225787fc19611b65cd57/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Disable when using a uclibc version before 0.9.33 since dn_comp
function support was added in this version. Also disabling on AVR32
since any AVR32 toolchain will be based on a uclibc version older
than 0.9.33 (for using an external AVR32 toolchain).
[Peter: use dn_comp instead of __dn_comp]
Signed-off-by: Ryan Barnett <rjbarnet@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since quite a while, we have deprecated, and then removed the support
to build a compiler on the target. Therefore, having a distcc package
for the package is quite useless, and this patch consequently marks it
as deprecated so it can be removed in a future version of Buildroot.
Fixes:
http://autobuild.buildroot.net/results/16b/16be2138c8e5ba785fa2ad55b478dcd1b6fb5123/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
mpg123 needs MMU because the application that is built with this
package uses fork. Currently it is difficult to only build the
libraries for mpg123 so disabling the package all together when there
is no MMU support.
Note: mpg123 is an optional dependency of mpd but mpd already requires
BR2_USE_MMU so there is no need to add this as a dependency.
Fixes:
http://autobuild.buildroot.net/results/5b0/5b053af566dd122ae7e58893e77d5d5f3070fb9e
Signed-off-by: Ryan Barnett <rjbarnet@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
In commit 87492d244a ("aiccu: disable
for uClibc 0.9.31/0.9.32"), we made sure it was not possible to select
aiccu with uClibc 0.9.31 and 0.9.32, because they lack dn_skipname().
However, we still have the problem that external AVR32 toolchains can
select aiccu, which causes build failures. Therefore, we also disallow
aiccu on AVR32 altogether. We keep the 0.9.31/0.9.32 exclusions,
because if they are used on other architectures, it would also fail
due to the lack of dn_skipname().
Fixes:
http://autobuild.buildroot.org/results/a24/a2490d434152625d9208615d83f4c5d6daea79d0/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The libeina library uses the madvise() system call, that isn't
available on non-MMU systems. Also, several other components of EFL
use fork(). Therefore, the easiest solution is to simply disallow the
EFL as a whole on non-MMU systems.
Fixes:
http://autobuild.buildroot.org/results/ad9/ad90baa5e07569308a7e2b2510b67c5b2a563b44//
Thanks to Ryan Barnett for helping in the investigation!
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Evas has an optional mechanism to do asynchronous preloading of
images. This mechanism is optional, and in commit
b6d92bf415 ("libevas: async image
preload support needs threads support in toolchain"), Peter made sure
to disable the asychronous preloading when no thread support was
available.
Unfortunately, it seems like disabling the asynchronous loading is
rarely used, and it in facts fails to build: a member of structure is
not present when asynchronous preloading is disabled, but the code
continues to use it.
Since the fix is not obvious, and all this mechanism seems to have
changed completely in EFL 1.8.x, and we probably don't care much about
EFL without threads, this commit adds a dependency of libevas on
thread support. Consequently, it also reverts commit
b6d92bf415 which is no longer necessary.
Of course, this commit propagates this additional dependency to the
reverse dependencies of libevas.
Fixes:
http://autobuild.buildroot.org/results/6de/6de90018a9eeb9c495d15046a8b3270eb95a5550//http://autobuild.buildroot.org/results/693/693df99db4ab357b48d427be3a72f6d64dd53065//
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
To be sure that host-autoconf dependency is already built move the
call to autogen.sh from SDL_POST_PATCH_HOOKS to SDL_PRE_CONFIGURE_HOOKS.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Acked-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The cross-compilation test is based on the ability to run
a test program on the host, which is wrong.
If it runs, then the configure script concludes
that we're doing native compilation,
if it doesn't run, we're doing cross-compilation.
The configure script needs to be regenerated to fix the
cross-compilation test.
Fixes
http://autobuild.buildroot.net/results/969/969a49ae97a50634ea846a82b9c360e4fb020ace/build-end.log
Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The tvheadend configure script fails to correctly handle spaces in any of
the options passed to it. If ccache is used the configure script fails to
handle the space in the --cc option and errors. Also the --cflags option
fails to handle multiple entries and only uses the first one. This patch
simple removes setting the --cc and --cflags options as the configure
script will correctly use the CC and CFLAGS variables already passed in
TARGET_CONFIGURE_OPTS.
Signed-off-by: Martin Bark <martin@barkynet.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
AVR32 is using uClibc 0.9.31, which lacks some CAN related definitions
that prevent the gpsd nmea2000 driver from building.
The rest of gpsd, with all options and drivers enabled, builds fine on
AVR32.
Fixes:
http://autobuild.buildroot.org/results/ee2/ee2ec848e893f08fa80caf99a67e68b73b6400e8/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
coreutils was not building in the static link case with uClibc, due to
binaries being linked with librt (for timer functions), but not with
libpthread. However, librt uses functions from libpthread, so both
must be specified. This lack of linking against libpthread was caused
by a deficiency in one of the gnulib m4 macro, for which we add a patch.
Fixes:
http://autobuild.buildroot.org/results/bd4/bd4cc779b46e3837e3d6c43c97eaf42fdfc3a288/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
pv configure script is somewhat stupid: if it cannot find the host
gettext tool msgfmt, it concludes that gettext is not available, and
provides its own minimal version. Unfortunately, this minimal version
conflicts with the available target gettext. We fix this by ensuring
that host-gettext is built if gettext support is enabled, by adding it
to the dependencies.
Fixes:
http://autobuild.buildroot.org/results/d30/d30db79190f2cf8953751d55a5d34644f60daec4/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
AArch64 does not implement the SYS_open system call, but only the
SYS_openat system call, as should be done for all modern architectures
merged into the Linux kernel. This causes problems when building
libv4l 0.8.9, which are fixed by integrating an upstream patch,
available in a more recent version of libv4l. Backporting the patch is
preferred over bumping libv4l, as libv4l has a fairly large number of
reverse dependencies, and we don't want to break anything that close
to the release.
Note that this commit also renames the existing libv4l-largefile.patch
to libv4l-01-largefile.patch in order to comply with the Buildroot
patch naming scheme, and guarantee the order in which patches are
applied.
Fixes:
http://autobuild.buildroot.org/results/ed6/ed66792e332ee0256a17f98cbe21dfcfe6e8743b/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
In commit 52eeb04c76 ('libsepol: fix
static-only library build') I forgot to git add the changes made to
the libsepol.mk file itself. Those changes are needed to actually pass
STATIC=<something> when building static libraries only.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
icu does not recognize 'uclinux' as a supported platform. While making
it recognize uclinux is easy, there is another problem down the road:
icu does very weird things to generate an ELF library containing
static data (libicudata.a), and the generated library being ELF, it is
not compatible with the FLAT binary format expected by uclinux
platforms such as Blackfin in FLAT format.
Therefore, we simply disallow the selection of icu on FLAT
platforms.
Note that adding a dependency on BR2_BINFMT_ELF doesn't work, because
BR2_BINFMT_FDPIC is considered to be separate (even if technically
FDPIC is a derivative of ELF). That's why the dependency we're adding
is "depends on !BR2_BINFMT_FLAT" and not "depends on BR2_BINFMT_ELF".
Fixes:
http://autobuild.buildroot.org/results/b41/b415fed7fae4012bad7d8b53a481bd71bdab716f/build-end.log
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
webkit needs the ARM blx instruction, which is only available on >=
ARMv5t, so we have to make sure the webkit package cannot be enabled
on < ARMv5t. In order to achieve this, this commit introduces the
BR2_PACKAGE_WEBKIT_ARCH_SUPPORTS hidden Config.in knob, which avoids
duplicating all over the place the complex architecture dependencies
of webkit.
Fixes:
http://autobuild.buildroot.org/results/fdf/fdf8bc7660ac251792d0542d2729ea22753d3789/build-end.log
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes a number of issues. From the changelog:
- Avoid getting stuck in a loop writing huge key files, reported by Bruno
Thomsen
- Avoid crash on exit due to cleaned up keys before last packets are sent,
debugged by Ronald Wahl
- Fix a race condition in rekeying where Dropbear would exit if it received a
still-in-flight packet after initiating rekeying. Reported by Oliver Metz.
This is a longstanding bug but is triggered more easily since 2013.57
- Ensure that generated RSA keys are always exactly the length
requested. Previously Dropbear always generated N+16 or N+15 bit keys.
Thanks to Unit 193
- Fix DROPBEAR_CLI_IMMEDIATE_AUTH mode which saves a network round trip if the
first public key succeeds. Still not enabled by default, needs more
compatibility testing with other implementations.
- Fix signal handlers so that errno is saved, thanks to Erik Ahlén for a patch
and Mark Wickham for independently spotting the same problem.
And at the same time get rid of the TIMEVAL_TO_TIMESPEC as dropbear provides a
fallback implementation if not available (and has done so since 2006).
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
3.3.8 support is removed since 334dca61 (kernel-headers: remove deprecated
versions 3.1, 3.3, 3.5), so let's get rid of the patch as well.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Host version is needed to build xbmc
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Host version is needed to build xbmc
[Peter: use _PRE_CONFIGURE_HOOKS like target version]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Remove a leftover of the commit commit e543f5a104
(arch: remove sh2, sh3 and sh3eb support).
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The https://code.google.com/p/pyasn/ project is not really the real
upstream for PyASN, and at least not the upstream for the PyASN
implementation recommended by the PySNMP developers.
Instead, the real upstream is
https://pypi.python.org/packages/source/p/pyasn1/, which has had much
more regular releases than the other PyASN implementation.
Therefore, we switch to using this implementation, as recommended by
the PySNMP developers on http://pysnmp.sourceforge.net/download.html.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The python and python3 builds mark libpython as read-only which
prevents it from being stripped out correctly for the target.
Signed-off-by: Przemyslaw Wrzos <przemyslaw.wrzos@calyptech.com>
Acked-by: Thomas De Schampheleire <thomas.de_schampheleire@alcatel-lucent.com>
Tested-by: Thomas De Schampheleire <thomas.de_schampheleire@alcatel-lucent.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
[Thomas: bump to 2.1.2 instead of 0.8, remove comment that no longer
made sense about setuptools being forked.]
Signed-off-by: Rohan Fletcher <rohfledev@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As we are going to bump setuptools to a much newer version, the host
python needs to be built with support for unicodedata.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Until now, Python external modules were only visible when Python 2.x
was selected. With this commit, we now source all the Config.in files
of Python external modules, as soon as one of the two Python
interpreters is enabled.
Since all Python external modules have a "depends on
BR2_PACKAGE_PYTHON" in their Config.in, this commit in practice does
not allow to enable any Python external module. However, thanks to
this, we can progressively and safely enable more and more Python
external modules to build with Python 3, by simply changing their
dependency to "depends on BR2_PACKAGE_PYTHON || BR2_PACKAGE_PYTHON3".
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This commit improves the Python package infrastructure to allow Python
packages to be built with Python 3. The changes are fairly simple:
* Use either PYTHON_PATH or PYTHON3_PATH as the PYTHONPATH depending
on which Python is used.
* Depend on host-python or host-python3 and python or python3
depending on which Python is used.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The Python package infrastructure will need the Python 3 package to
provide a PYTHON3_PATH environment variable in order to build
third-party Python modules.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This commit bumps the Python3 package to use Python 3.4.0rc1.
About the patches:
* The patches below 100 are significantly changed, because like for
Python 2.x, a good number of improvements have been made in the
upstream Python for cross-compilation. Therefore, almost all of
these patches have been modified.
* All the patches above 100 are simply updated for Python 3.4.0, with
a small refactoring for the handling of test modules.
The details of the python3.mk changes are:
* --without-ensurepip to tell Python to not use PIP at build time.
* Many environment variables are no longer passed, they were specific
to our cross-compilation patches
* The fixup of the LIBDIR in the Python Makefile is no longer needed
since Python has switched to _sysconfigdata.py for distutils
configuration instead of parsing the Makefile.
* A new post patch hooks touches the two files generated by pgen to
make sure they are newer than the pgen sources, which ensures pgen
is not built/executed.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Some parts of python3.mk were hardcoding the 3.3 version as the major
version, which does not work for Python 3.4 and other future
versions. Instead, use the existing PYTHON3_VERSION_MAJOR.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The host python always had --disable-unicodedata, regardless of the
corresponding configuration option BR2_PACKAGE_PYTHON_UNICODEDATA.
Since the host python is used to byte-compile python modules, this meant
that such modules could not contain unicode strings. For example, following
statement in a python module:
print u"\N{SOLIDUS}"
would cause the byte-compilation to fail with message:
SyntaxError: ("(unicode error) \\N escapes not supported (can't load
unicodedata module)",
Instead, conditionally disable unicodedata based on
BR2_PACKAGE_PYTHON_UNICODEDATA, also for the host python.
This fixes bug #6542 (https://bugs.busybox.net/show_bug.cgi?id=6542)
Reported-by: Gernot Vormayr <gvormayr@gmail.com>
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The target python3 depends on host-python3, but most of the scripts
call "python", so we need to ensure that $(HOST_DIR)/usr/bin/python
exists. This patch achieves this by creating a python -> python3
symbolic link in $(HOST_DIR), just like we are already doing for the
target Python 3.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
In Buildroot, we do not support installing both Python 2.x and Python
3.x on the target.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Thanks to the previous commit that makes distutilscross unecessary for
setuptools packages, the host-distutilscross package can now be
removed. There is no need for any Config.in.legacy handling, since
there is no target variant, or visible Config.in option for
host-distutilscross.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Thanks to the bump of Python 2.x, distutilscross is no longer needed
to achieve cross-compilation for setuptools packages. The host Python
2.x interpreter can be tricked into using the target compiler thanks
to pointing it to a different sysconfigdata module, which is achieved
using PYTHON_PATH.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Transmission would incorrectly determine the C++ compiler when ccache is
enabled, causing a build with uTP to fail at the configure step.
This patch adds a patch against transmission, fixing the problem.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
If BR2_PACKAGE_QT_STATIC is set, qtuio will not build a .so file, but .a.
However, the custom INSTALL_TARGET_CMDS and INSTALL_STAGING_CMDS
unconditionally attempted to copy the .so file.
This commit checks the requested Qt library type and copies the right
library for each case, taking into account that the static .a file does not
need to be copied to the target directory.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
If Lua support is requested in VLC, its configure script tries to find the
luac byte compiler, which fails if host-lua is not yet built.
This can be easily reproduced by setting a minimal config with vlc and Lua
enabled, and running: 'make clean toolchain vlc'. Final output is:
checking for LUA... no
configure: WARNING: Package lua5.2 was not found in the pkg-config search path.
Perhaps you should add the directory containing `lua5.2.pc'
to the PKG_CONFIG_PATH environment variable
No package 'lua5.2' found, trying lua 5.1 instead
checking for LUA... no
configure: WARNING: Package lua5.1 was not found in the pkg-config search path.
Perhaps you should add the directory containing `lua5.1.pc'
to the PKG_CONFIG_PATH environment variable
No package 'lua5.1' found, trying lua >= 5.1 instead
checking for LUA... yes
checking for luac... no
configure: error: Could not find the LUA byte compiler.
make: *** [<buildroot>/output/build/vlc-2.1.2/.stamp_configured] Error 1
Fix this problem by setting host-lua as a dependency to vlc.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
It does not make much sense enabling opencv without its core module.
This configuration leads to build nothing (since all modules depend on
the core one), but install the configuration files (*.pc and *.cmake)
anyway.
This absurd situation may break the build-system of other packages
that would correctly find the *.pc (but does not check for the modules
they actually use), but would not build because of missing headers and
libraries.
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
jimtcl tries to use 'ccache' (a non-buildroot host version) which may not
exist on the host system. If ccache is enabled in buildroot, the compiler
used by jimtcl is:
ccache <buildroot>/.../ccache <buildroot>/.../<tuple>-gcc
If ccache is not present on the host, this results in the build error:
ccache <buildroot>/.../ccache <buildroot>/.../<tuple>-gcc
-D_GNU_SOURCE -Wall -I. -fpic -pipe -Os -c -o jim-subcmd.o jim-subcmd.c
make[1]: ccache: Command not found
This patch passes 'CCACHE=none' to the 'configure' script, disabling the
internal handling of ccache, so that ccache can be transparently passed
through CC.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The expanded SED variable already contains -e, so the extra -e was being
interpreted as the sed command and causing sed to fail.
Signed-off-by: Ivan Sergeev <vsergeev@kumunetworks.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Thanks to the Python 2.x bump, it is no longer needed to pass
PYTHONCPREFIX, and CROSS_COMPILING when building third-party Python
modules.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Even though jumping from 2.7.3 to 2.7.6 looks like a minor version
bump, it is in fact a fairly significant one, because a good number of
changes to help cross-compilation have been merged into Python
upstream. Therefore, most of our patches are affected by this change.
In detail, this commit:
* Renames all the patches to follow the naming convention of patches
in Buildroot: the patch file names should not have any version
number.
* The patches numbered above 100, that add configuration options to
disable certain modules of the Python standard library, are only
renamed and slightly adapted, they didn't change that much.
* The patches numbered below 100 are almost entirely rewritten: many
of the cross-compilation problems that used to exist in Python
2.7.3 no longer exist, and the number of remaining problems is
smaller, and can be fixed with smaller patches.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
With the upcoming bump of Python 2.x, it will become important that
the PYTHONPATH is passed whenever we build third-party packages, be
they using the distutils build mechanism, or the setuptools build
mechanism. This is because passing PYTHONPATH is what will allow
Python to find a special Python module that contains all the
compiler/library/headers definitions that are relevant when
cross-compiling.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Now that the Python package exposes its PYTHON_PATH variable, we can
use it in the package infrastructure. This prepares both the upcoming
bump of Python 2.x, and the introduction of Python 3 support in the
Python package infrastructure.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As a preparation to make the Python infrastructure support both Python
and Python 3, as well as the bump of Python 2 and 3, we need the
Python package to expose the Python module path in a variable called
PYTHON_PATH. It will be used by the following commits.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The three typical packages that use .config files in buildroot copy the
config file at different times in the build process:
busybox copies its .config from the post-extract hook.
linux copies its .config in the configure_cmds.
uclibc copies its .config from the post-patch hook.
Copying the .config file from the configure step is the only way to properly
support an OVERRIDE_SRCDIR that does not yet have the .config file, because
the extract and patch steps are skipped in that case.
In a previous patch, the situation was already fixed for busybox. This patch
applies the same fix to uclibc: copy the config file from the configure
step, as linux is doing.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The three typical packages that use .config files in buildroot copy the
config file at different times in the build process:
busybox copies its .config from the post-extract hook.
linux copies its .config in the configure_cmds.
uclibc copies its .config from the post-patch hook.
Copying the .config file from the configure step is the only way to properly
support an OVERRIDE_SRCDIR that does not yet have the .config file, because
the extract and patch steps are skipped in that case.
For example, when setting a BUSYBOX_OVERRIDE_SRCDIR to a cleanly extracted
busybox tarball:
$ make busybox-dirclean busybox
rm -Rf [..]/output/build/busybox-custom
>>> busybox custom Syncing from source dir
>>> /home/tdescham/repo/contrib/busybox-1.21.1
rsync -au --exclude .svn --exclude .git --exclude .hg --exclude .bzr
--exclude CVS /home/tdescham/repo/contrib/busybox-1.21.1/
[..]/output/build/busybox-custom
>>> busybox custom Configuring
/bin/sed -i -e "/\\<CONFIG_NOMMU\\>/d"
[..]/output/build/busybox-custom/.config
/bin/sed: can't read [..]/output/build/busybox-custom/.config:
No such file or directory
make: *** [[..]/output/build/busybox-custom/.stamp_configured] Error 2
This patch modifies busybox.mk to copy the config file from the configure
step instead, as linux is doing, and fixing the described scenario.
This fixes bug #5030: https://bugs.busybox.net/show_bug.cgi?id=5030
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This patch consolidates the URLs for various Freescale-supplied
packages to use FREESCALE_IMX_SITE.
Signed-off-by: Eric Nelson <eric.nelson@boundarydevices.com>
Reviewed-by: "Thomas Petazzoni" <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
To be sure that host-autoconf dependency is already built move the
call to autogen.sh from SDL_POST_PATCH_HOOKS to SDL_PRE_CONFIGURE_HOOKS.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Acked-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Add '+' prefix to the $($(PKG)_BUILD_CMDS) and $($(PKG)_INSTALL*_CMDS)
commands to enable jobserver for the sub-make.
Without the '+' prefix GNU make does not detect the sub-make so it
disable the jobserver for the sub-make.
>From GNU make documentation:
Using the MAKE variable has the same effect as using a ‘+’ character
at the beginning of the recipe line. This special feature is only
enabled if the MAKE variable appears directly in the recipe: it does
not apply if the MAKE variable is referenced through expansion of
another variable. In the latter case you must use the ‘+’ token to get
these special effects.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
To be able to use top-level parallel make we must not depend in a rule
on the order of evaluation of the prerequisites, so instead of relying
on the left to right ordering of evaluation of the prerequisites add
an explicit rule to describe the dependencies.
We cannot use the pattern rules because they must have the same
dependency for every package, but we need to change the dependencies
depending on $(2)_OVERRIDE_SRCDIR variable value, so we must use a
more flexible way like $(2)_TARGET_% variables.
So add explicit dependencies for the following stamp files:
$(2)_TARGET_EXTRACT
$(2)_TARGET_PATCH
$(2)_TARGET_CONFIGURE
$(2)_TARGET_BUILD
$(2)_TARGET_INSTALL_STAGING
$(2)_TARGET_INSTALL_TARGET
$(2)_TARGET_INSTALL_IMAGES
$(2)_TARGET_INSTALL_HOST
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This commit makes the dependency from the target toolchain explicit.
This way we can buid from command line a package that use
inner-generic-package right after the configuration phase, example:
make clean <package-name>
Also remove TARGETS_ALL because the only purpose was to add toolchain
dependency so it's superseded by this commit.
To prevent circular dependency add the new variable
<pkgname>_ADD_TOOLCHAIN_DEPENDENCY to avoid adding the toolchain
dependency for toolchain packages.
This is also a step forward supporting top-level parallel make.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Move "dependencies" "dirs" "prepare" dependencies from "toolchain" to
every package.
This way we can build correctly every package right after the clean
stage.
As example with this commit we can build successfully the glibc right
after the clean stage:
make clean glibc
This is also a step forward supporting top-level parallel make.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since we are changing the default dbus user, make sure this user is
consistently created.
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
[Peter: add to existing _CONF_OPT line as suggested by Yann]
Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Change the download source to a Github repository which is more recent.
This patch also bump libatomic_ops version and remove a patch that is
now upstream.
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
glibc 2.19 has been released recently
(https://sourceware.org/ml/libc-alpha/2014-02/msg00224.html). This
commit allows to build a toolchain with this new version. In order to
allow this, we add a version selection that did not exist for
glibc. We default to 2.18, which was the only supported version until
now, and add an option for 2.19.
For microblaze, which uses a specific glibc version, the version
selection choice is not displayed.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This reverts commit 262a4c0bf7.
Compiler error has been fixed, and building this package doesn't cause an
ICE anymore.
Signed-off-by: Anton Kolesov <anton.kolesov@synopsys.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, the build of systemd is broken with the internal toolchain
backend, because we have uClibc patches that add support for
execvpe(), but we also have a patch for systemd that adds execvpe(),
which was added when the internal uClibc didn't support execvpe().
However, simply dropping the patch is not a solution, as it would
break the build of systemd with any other uClibc than the ones built
with Buildroot. For example, a Crosstool-NG uClibc toolchain would
fail.
Beyond this execvpe() problem, there are also other problems that
prevented systemd from being built with unpatched uClibc.
This patch does the following:
* Add sequence numbers of systemd patches.
* Regenerate them to avoid fuzz.
* Improve the uClibc compatibility patch so that it:
- Detects whether execvpe() is available or not, and if not
available provide an implementation.
- Detects if dup3() is available or not, and if not available
provide an implementation.
- Detects if MSG_CMSG_CLOEXEC is defined or not, and if not, define
it to the appropriate value.
This has been tested with uClibc internal toolchain, uClibc external
toolchain built with Buildroot, uClibc external toolchain built with
Crosstool-NG, and glibc external toolchain.
Fixes bug #6776.
Fixes:
http://autobuild.buildroot.org/results/591/591c6055430da334bd1e46e7d01497add45da837/build-end.loghttp://autobuild.buildroot.org/results/e58/e5851939d2837d4e35a4c1d9ca6df3c93ed5a34b/build-end.loghttp://autobuild.buildroot.org/results/92e/92e4c40c69a3feb4046b2fe5a0d8c69d5a44a157/build-end.log
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This patch bumps systemd to v207 but also declares it as a provider for the
udev virtual package.
Starting with systemd 183, udev has been merged into
systemd. The udev daemon is now installed as /lib/systemd/systemd-udevd.
This means that /dev management using udev is only available if systemd
is chosen as init system.
When configuring systemd, the following options are available:
- activation of systemd-journal-gatewayd, to access the journal via
HTTP.
- activation of extra features like journal compression and sealing.
Support for uClibc has also been removed because:
- upstream has no interest in supporting uClibc.
- using a shrinked libc brings no advantage, given the size of all the
programs included in Systemd. So using glibc does not matter.
Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This patch converts udev to a virtual package. For the moment, there is only
one provider for the udev features: eudev.
Packages meant to provide udev-like features must select the symbol
BR2_PACKAGE_HAS_UDEV.
Packages depending on BR2_ROOTFS_DEVICE_CREATION_DYNAMIC_UDEV or
BR2_PACKAGE_UDEV have been converted to use the new symbol.
[Peter: move legacy symbols under 2014.05]
Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
eudev is a userspace device management daemon. It is a standalone
version, independent from systemd. It is a fork maintained by Gentoo.
Features:
- No extra configuration options are available: Gudev is build if
libglib2 is selected.
- No dependency on hwdata as the package uses its own hardware
database (as does systemd).
eudev 1.3 is in sync with systemd v207.
[Peter: add BR2_USE_MMU dependency]
Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As the e-mail address buildroot@buildroot.org is now enabled, update the
e-mail addresses in the source tree from @uclibc.org and @busybox.net to our
own proper domain.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Disable elfutils for avr32, because:
- elfutils uses canonicalize_file_name, which came after uClibc-0.9.31.
- avr32 uses a toolchain with uClibc-0.9.31
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Adjust phpize and php-config to make them work for cross-compiled
external extensions.
While at it also fix dl* issues that prevent said extensions from
loading.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The patch that commit a728e2fe35 (coreutils: fix build against uclibc
snapshot) adds mentions the wrong coreutils version in its description.
Also fix the sign-of tag while at it.
Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The xtensa port uses __xtensa_libgcc_window_spill in libgcc to implement
__builtin_frame_address. This symbol is local/hidden in libgcc. This is not a
problem when linking against static libgcc. But g++ defaults to
-shared-libgcc, thus breaking link against C++ shared libraries that are using
__builtin_frame_address as follows:
ld: test: hidden symbol `__xtensa_libgcc_window_spill' in .../libgcc.a(lib2funcs.o) is referenced by DSO
Add upstream patches that make libgcc_s.so a linker script that links in
unresolved symbols from the static libgcc, similar to the ARM and PowerPC
ports.
Fixes:
http://autobuild.buildroot.net/results/e2d/e2d1a763fa86b8575e2e48e6d73c018175f43e7c/
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
uClibc development version adds support for POSIX spawn routines. However,
unlike glibc these routines are in librt. This breaks gnulib autoconf
detection. Teach gnulib autoconf to look for POSIX spawn in librt.
Fixes:
http://autobuild.buildroot.net/results/246/246b3778a1a646afd1c8b9c17b4579fb5a27120e/
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As reported on IRC, our default busybox configuration doesn't support SHA
encoded passwords, breaking login if enabled under system configuration.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
20140129:
Minor bug fix in configs 19d2:0149 and 19d2:2000 (Lars Melin reported)
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Version 2.1.0, 2014/01/28
ATTENTION: -I flag meaning reversed, default is to skip SCSI inquiry;
introduction of StandardEject, replacing many MessageContents with the
same function, reducing size of device config files, and always including
the 'Allow Medium Removal' before ejecting (thanks to Lars Melin for
the idea); fix in "bulk_read", removing bogus CSW request (report from
"Sonya@zte")
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This version includes:
- add f2fstat tool
- critical bug fix on nat bitmaps
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Context:
The autobuilders were failing on the symbol _XData32 being in conflicts.
A patch had been added to SDL to add a check to the configure.in
Problem:
Sometimes, the build would fail, because of an _XData32 symbol being in
conflicts eventhrough the patch was here.
What was happening:
Following the classic buildroot workflow:
- Extract
- [...]
- Apply 001 patch, which touches configure.in AND configure
- Apply 002 patch, which touches configure.in
- Invoke autogen.sh
- [...]
Right before running autogen.sh, we have configure.in which is more
recent than configure, which is fine.
We then, execute autogen.sh which, basically, runs autoconf.
If your machine was lighty loaded, the time difference between
configure.in and configure was really tiny (ms order), which seems to be
neglected by autoconf.
The results was that the configure was *NOT* generated. And our second
patch was not taken into account.
If your machine was under heavy load, the time difference between the
two files would have been greater and then *maybe* picked up by
autoconf. And then the configure file was re-generated.
When the 0001 patch was introduced, SDL package did *NOT* run it's
autogen.sh, which is why it touches also the configure.
This came later, causing this behavior.
Fixes:
http://autobuild.buildroot.net/results/d1c/d1c36f634dbf6b6e5d18444c2a23dfd129202b80/
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
