Saul Wold [Fri, 11 Jan 2013 23:03:54 +0000 (15:03 -0800)]
xz: LICENSE field is wrong
xz contains a GPLv3 m4 macro which is used at build time, this does not make
the installed or runtime items GPLv3, which is what the LICESNE field pertains
to, therefor remove it from the LICENSE field.
Emilia Ciobanu [Mon, 14 Jan 2013 09:04:33 +0000 (11:04 +0200)]
distrodata: Updated checkpkg task
bitbake -c checkpkg fetches the latest version for all packages,
including the ones from sourceforge and those that have
ununsual package naming.
Also removed all pn-native/nativesdk-pn entries from the resulting
checkpkg.csv file if the system already reported the latest version
for pn.
Eric Bénard [Fri, 11 Jan 2013 13:51:25 +0000 (14:51 +0100)]
add meta-toolchain-qt
- this allows the creation of a SDK containing all the needed
libraries to develop QT4 for X applications
- the recipes were factorized to share as much code as possible
- binaries generated using the toolchain were run tested on an
ARMv7 target
- both qt & qte toolchains were build tested for qemuarm target
and for an ARMv7 target
Signed-off-by: Eric Bénard <eric@eukrea.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Marko Lindqvist [Thu, 10 Jan 2013 09:59:55 +0000 (11:59 +0200)]
python: replace obsolete automake macros with working ones
Add obsolete_automake_macros.patch for both python-dbus and
python-pygobject that replace automake macros no longer supported
by automake-1.13 with modern constructs.
Signed-off-by: Marko Lindqvist <cazfi74@gmail.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Marko Lindqvist [Thu, 10 Jan 2013 18:58:01 +0000 (20:58 +0200)]
coreutils: fix license segment md5sum boundary
Replace incorrect "startline" with correct "beginline" for telling
where license segment in file begins. Old md5sum was calculated
from the beginning of the file, not beginning of the license segment.
Signed-off-by: Marko Lindqvist <cazfi74@gmail.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Bruce Ashfield [Thu, 10 Jan 2013 20:49:16 +0000 (15:49 -0500)]
linux-libc-headers: fix headers install in long path name environments
If a build is started in a deep directory structure, or in a path with
long directory names the installation of linux-libc-headers will fail
with:
| make[2]: execvp: /bin/sh: Argument list too long
| make[2]: ***
>From within the kernel Makefiles themselves.
The solution is to patch the kernel build rules with the following
change:
[
scripts/Makefile.headersinst: install headers from scratch file
If headers_install is executed from a deep/long directory structure, the
shell's maximum argument length can be execeeded, which breaks the operation
with:
| make[2]: execvp: /bin/sh: Argument list too long
| make[2]: ***
By dumping the input files to a scratch file and using xargs to read the
input list from the scratch file, we can avoid blowing out the maximum
argument size and install headers in a long path name environment.
]
Until this change, or a similar one, is merged into the mainline kernel, this
change should be applied to any 3.7 or greater libc-headers build.
Upstream-status: Pending
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 9 Jan 2013 16:48:08 +0000 (16:48 +0000)]
gtk+3: add 3.4.4
Mostly taken from meta-gnome, with some changes.
Instead of depending on gtk+-native, depend on just gdk-pixbuf-native and build
the tool required for the host directly. This saves building the entire GTK+
stack natively just for one tool.
Append a version suffix and use update-alternatives on gtk-update-icon-cache as
with GTK+ 2.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 9 Jan 2013 16:48:06 +0000 (16:48 +0000)]
gtk+: version installed gtk-update-icon-cache and use alternatives
With GTK+ 3 being added soon we'll have potentially two providers of
gtk-update-icon-cache. Append a version to the binary and use
update-alternatives to ensure that the unversioned name works.
For gtk+-native the alternatives won't take effect, so install a symlink
explicitly. This will break if we have a gtk+3-native but we can fix that if
and when that happens.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Samuel Stirtzel [Thu, 10 Jan 2013 10:35:47 +0000 (11:35 +0100)]
boost: Use PARALLEL_MAKE for bjam
Like make, bjam accepts the parameter -jX for multithreaded execution
do_install also profits from this setting
Tested with a quad core 64bit intel cpu (with hyperthreading) PARALLEL_MAKE="j16"
$ time bitbake boost
before:
real 14m37.433s
user 7m40.785s
sys 4m30.109s
after:
real 7m11.979s
user 12m10.694s
sys 2m47.078s
Also fixes tab indention
Signed-off-by: Samuel Stirtzel <s.stirtzel@googlemail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Florin Sarbu [Thu, 10 Jan 2013 07:01:09 +0000 (09:01 +0200)]
alsa-utils: add bash as a dependency of the alsa-utils-alsaconf rpm.
alsa-utils-alsaconf rpm contains the bash script alsaconf. In order
to have the do_rootfs not error when bash is not brought along in
the rootfs as an explicit dependency, we add it as a dependency
of alsa-utils.
Signed-off-by: Florin Sarbu <florin.sarbu@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Hongxu Jia [Thu, 10 Jan 2013 07:58:10 +0000 (15:58 +0800)]
apt-native: fix the creation of apt.conf.sample
1.The file of "apt.conf.sample" is in the outdir, and outdir is assigned
by "os.path.join" with the params of ${D}, ${sysconfdir} and "apt". But
${sysconfdir} is an absolute dir and that let ${D} be discarded.
The following is the help on function os.path.join(a, *p):
Join two or more pathname components, inserting '/' as needed.
If any component is an absolute path, all previous path components
will be discarded.
Use oe.path.join instead which don't have this problem.
2. Create apt.conf.sample if it doesn't exist.
[YOCTO #3677]
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Marko Lindqvist [Wed, 9 Jan 2013 12:13:53 +0000 (14:13 +0200)]
harfbuzz: add recipe, version 0.9.10
OpenType text shaping engine HarfBuzz is dependency of pango >= 1.32
This depends on all of glib-2.0, cairo, and freetype to consistently
provide all backends. Otherwise it would, in typical case, depend
on build order of these packages. HarfBuzz configure would pick those
backends for which dependencies happen to be built before it.
Similarly graphite is consistenly disabled if there ever happens
to be recipe for it.
Signed-off-by: Marko Lindqvist <cazfi74@gmail.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Khem Raj [Sat, 5 Jan 2013 03:19:22 +0000 (19:19 -0800)]
udev: Fix build with eglibc 2.17
Link with librt to get definitions of mq_* functions
Fixes errors like below
/home/kraj/work/angstrom/sources/openembedded-core/build/tmp-eglibc/work/armv5te-oe-linux-gnueabi/udev/182-r3/udev-182/src/sd-daemon.c:394:
undefined reference to `mq_getattr'
| collect2: error: ld returned 1 exit status
Laurentiu Palcu [Thu, 20 Dec 2012 13:15:39 +0000 (15:15 +0200)]
gtk-immodules-cache.bbclass: allow for offline cache generation
In order to support a RO rootfs, the cache generation during postinstall
has to be done on host. However, gtk-query-immodules application will
only be able to parse shared objects from the same ELF class.
In order not to have a native package for all the recipes providing
an input method, so we can generate the cache file natively, run
gtk-query-immodules through qemu emulator.
kernel.bbclass: Define MODULE_IMAGE_BASE_NAME for layer convenience.
Define the additional variable MODULE_IMAGE_BASE_NAME since some
layers (like meta-ti) would find it handy for multi-kernel
definitions in that they wouldn't have to define their own
variable representing exactly the same thing.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Ross Burton [Wed, 9 Jan 2013 16:26:04 +0000 (16:26 +0000)]
musicbrainz: handle rebuilds
When rebuilding build-native already exists. Sadly handling the directory
existing causes cmake to forkbomb when rebuilding, so delete an existing
build-native.
Also remove a variable that cmake warns isn't being used, and just build the
tool that we need natively, not the entire library.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Otavio Salvador [Tue, 8 Jan 2013 14:19:39 +0000 (12:19 -0200)]
libnl: Update to 3.2.18
This release fixes the build failures found when building against old
Linux kernel headers and also includes the fix for to use 'flex'
instead of 'lex'.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Saul Wold <sgw@linux.intel.com>
The license file text was updated to extend the copyright in 2011,
to reflect the new package version and to fix various typos. Tested
by compiling and running guile (which depends on bdwgc) on qemu-x86.
lttng2-ust, lttng-modules and lttng-tools were updated to version
2.1.0. Since these should be changed together, the changes are
grouped into a single patch.
The text for the lttng-tools LICENSE file has these lines changed
(according to the license text, these shouldn't impact the actual
licensing terms of the package):
We have some tests LGPL but should not impact anything even if we change them
in the future to GPL.
=================================================================
The text for the lttng-modules LICENSE files contains a minor change in the
MIT section (a reference to the mit-license.txt file):
=================================================================
These files are licensed under an MIT-style license. See mit-license.txt
for details.
=================================================================
Marco Cavallini [Tue, 8 Jan 2013 07:51:44 +0000 (08:51 +0100)]
kernel.bbclass: kernel_do_install fails with kernel <2.6.30
* kernels <2.6.30 don't have $kerneldir/tools directory
* so we got this error:
* sed: can't read ...SNIP...linux/2.6.30-r0/image/usr/src/kernel/tools/perf/Makefile: No such file or directory
* ERROR: Function failed: do_install
Signed-off-by: Marco Cavallini <m.cavallini@koansoftware.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Cristian Iorga [Mon, 7 Jan 2013 14:24:59 +0000 (16:24 +0200)]
libtelepathy: python-native is now used during build
libtelepathy generated makefiles contain an
assignment like this: PYTHON=python
This could lead to potential build problems
in Yocto environment because python-native
is not used during the build.
Hongxu Jia [Wed, 9 Jan 2013 09:07:46 +0000 (17:07 +0800)]
sstate.bbclass:specify function dirs to avoid race
Specify dirs in which the shell function "sstate_create_package" and
"sstate_unpack_package" are executed and don't use ${B} as default dirs
to avoid possible race with task do_rootfs at deb image creation time.
[YOCTO #3674]
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Lukas Bulwahn [Wed, 9 Jan 2013 07:20:21 +0000 (08:20 +0100)]
site/common: deactivate a runtime check for ipv6-support in python >=2.7.1 that fails when cross-compiling
In configure there is a runtime check to determine buggy getaddrinfo and as we
are cross-compiling this check goes wrong. If ipv6 is enabled in python, the
failed test is reported as fatal error with the message
Fatal: You must get working getaddrinfo() function.
Setting ac_cv_buggy_getaddrinfo=no deactivates the runtime check and allows to
compile python with ipv6 enabled.
Chen Qi [Wed, 9 Jan 2013 02:31:57 +0000 (10:31 +0800)]
runqemu-internal: add 'debugshell' as a default kernel option
Add 'debugshell' as a default kernel option for ramfs booting.
If rootfs.img cannot be found under /media, init-live.sh loops
forever without showing any information. Silently looping forever
and blocking users is inappropriate.
Now that the 'debugshell' feature has been implemented in init-live.sh,
It's reasonable to add it to the kernel option when booting a ramfs-based
image. In this way, the system doesn't loop forever and instead drops
to a shell after a default timeout (30 seconds).
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Stan Hu [Wed, 9 Jan 2013 02:37:03 +0000 (19:37 -0700)]
wpa-supplicant: Include wpa_supplicant.service file in package for dbus to work
The dbus service files include a reference to wpa_supplicant.service, but if it
does not exist the dbus-daemon warns:
[system] Activating via systemd: service name='fi.w1.wpa_supplicant1' unit='wpa_supplicant.service'
[system] Activation via systemd failed for unit 'wpa_supplicant.service': Unit wpa_supplicant.service failed to load: No such file or directory. See system logs and 'systemctl status wpa_supplicant.service' for details.
If this happens, wpa_supplicant is not run automatically.
Signed-off-by: Stan Hu <stanhu@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Mon, 7 Jan 2013 15:43:06 +0000 (10:43 -0500)]
linux-libc-headers: make 3.7 the default and remove older libc-header recipes
Switch the default libc-headers to the 3.7 version. At the same time, remove
older versions of the headers to keep things simple and clear. All userspace
and kernel combinations should build and boot against this single lib-headers
version.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org> Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Eric Bénard [Mon, 7 Jan 2013 17:06:57 +0000 (18:06 +0100)]
qt4: blacklist untrusted SSL certificates
- this blacklist wrong certificates
https://bugreports.qt-project.org/browse/QTBUG-24654
https://bugreports.qt-project.org/browse/QTBUG-28937
- these patches will be in the next 4.8.5 release
Signed-off-by: Eric Bénard <eric@eukrea.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
