are not generally useful on target systems. Also, their presence in the
same binary package as libharfbuzz.so defeats the debian package autonamer.
Inherit lib_package to move these to ${PN}-bin.
Signed-off-by: Phil Blundell <pb@pbcl.net> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 14 Jan 2014 15:19:26 +0000 (15:19 +0000)]
cmake: specify all install paths
Specify the full set of install paths (bindir, libdir, etc) for packages that
use the GNUInstallDirs module, instead of just the prefix and leaving the rest
as default (which breaks with multilib).
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Mon, 13 Jan 2014 17:26:51 +0000 (17:26 +0000)]
grub-efi: remove configure patch, simply seed the cache
grub-2.00-disable-help2man patches configure.ac to disable the help2man check
because we don't need the man pages. It also then patches configure itself so
that autogen.sh doesn't have to be called for performance reasons. However,
do_configure causes a full autoreconf so this optimisation is moot, and can
cause patch failures when an existing build tree is re-used.
Instead, simply use CACHED_CONFIGUREVARS to tell configure that it can't find
help2man.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Robert Yang [Tue, 14 Jan 2014 02:42:13 +0000 (10:42 +0800)]
alsa-tools: fix the AUTOMAKE_DIR
There should be only one automake under the STAGING_DATADIR_NATIVE in
theory, but der_steffi@gmx.de has reported an odd problem which seemed
like that there are more than one. However, the "automake
--print-libdir" is the regular way to locate automake libdir.
[YOCTO #5706]
Reported-by: der_steffi@gmx.de Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Tue, 14 Jan 2014 11:43:14 +0000 (11:43 +0000)]
ptest: d.keys() is slow, use a list instead
Unfortunately d.keys is extremely slow. Using a list in this case should be
fine since the addtask lines are immediately above the code and aren't
going to change often.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Tue, 14 Jan 2014 11:42:38 +0000 (11:42 +0000)]
eglibc-locale: Fix depends on binutils
The dependency here needs to apply for nativesdk as well as target packages
as the autobuilder just tripped over that. We'd never want a native version
so I'm not sure why the target class override was even present. The dependency
also applies to do_package so lets be explicit about that in case sstate
decides to get clever.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 3 Dec 2013 15:42:58 +0000 (15:42 +0000)]
cmake: respect ${S} and ${B}
Instead of the class-specific variables OECMAKE_BUILDPATH and
OECMAKE_SOURCEPATH, just use ${B} and ${S}.
If these two paths are different, delete any existing ${B} before running a
build so that previous builds don't taint the current build.
Note that OECMAKE_SOURCEPATH and OECMAKE_BUILDPATH are not respected, so recipes
that manually set these in the past will need to be updated to either use
something along the lines of separatebuilddir.inc or set B themselves. If the
old variables are set, a warning is displayed.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Paul Eggleton [Tue, 17 Dec 2013 12:04:15 +0000 (12:04 +0000)]
dropbear: upgrade to 2013.62
LIC_FILES_CHKSUM has changed with the introduction of a BSD-3-Clause
algorithm (curve25519-donna); this has prompted a re-evaluation of the
LICENSE value which should now reflect the licenses declared in the
upstream documentation. Thanks to Beth Flanagan for helping with this.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Nathan Rossi [Fri, 10 Jan 2014 08:01:33 +0000 (18:01 +1000)]
ptest.bblass: Fix package QA issues when disabled
When the ptest distro feature is disabled, a ptest directory is still
created in the install phase, This directory is not cleaned up or
consumed by any package and will throw a QA error, e.g.
ERROR: QA Issue: glib-2.0: Files/directories were installed but not
shipped
/usr/lib/glib-2.0/ptest
ERROR: QA run found fatal errors. Please consider fixing them.
ERROR: Function failed: do_package_qa
This is caused by the do_install_ptest_base[cleandirs] attribute which
is not setup to be conditional on ptest being enabled.
This patch refactors the use of PTEST_ENABLED in the *ptest_base tasks,
replacing the conditional execution with the removal of the tasks from
the build, this prevents any part (including cleandirs) of the ptest
tasks from executing when disabled.
Ross Burton [Thu, 9 Jan 2014 16:01:51 +0000 (16:01 +0000)]
oeqa/runtime/systemd: wait for services to start/fail
When checking that no services have failed to start, actually wait for services
to finish starting by waiting for there not be no units in the "activating"
state.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Anders Darander [Fri, 10 Jan 2014 14:59:01 +0000 (15:59 +0100)]
terminal.bbclass: do not export PS1
With a complex PS1 setup, PS1 might not have all characters correctly escaped
when terminal.bbclass writes the export. This caused the run.do_terminal.PID to
terminate, making it impossible to use the devshell.
As the spawned shell will parse e.g. .bashrc (or whatever rc-file is being
used), PS1 will be reset in the devshell.
Signed-off-by: Anders Darander <anders@chargestorm.se> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Phil Blundell [Fri, 10 Jan 2014 14:38:32 +0000 (14:38 +0000)]
package: Keep global pkgfiles[] up to date when snapping library symlinks
Since ca86603607a69a17cc5540d69de0e242b33382d3 we are now calling stat() on
all the pkgfiles[] during emit_pkgdata(). If symlink snapping has removed
some of the files then we will blow up trying to stat a path that no longer
exists. Fix that by ensuring that pkgfiles[] is updated when we process the
list of library renames.
Signed-off-by: Phil Blundell <pb@pbcl.net> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Phil Blundell [Fri, 10 Jan 2014 12:54:16 +0000 (12:54 +0000)]
binutils: Also add autoconf-native to DEPENDS
Commit 616354f13732d13c17434d5b60b166f691c25761 is insufficient because
gnu-config-native's gnu-configize script uses perl modules from autoconf
and hence doesn't work unless autoconf-native is staged (which it may
not be if building from sstate).
Ideally g-c-n would itself declare a dependency on autoconf-native but this
is difficult to arrange without creating a dependency loop. autoconf-native
already depends on gnu-config-native (because autoreconf invokes gnu-configize)
and has a build dependency on m4-native, which in turn build-depends on g-c-n
because it configizes itself by steam in do_configure and needs config.{guess,sub}
to be available. Adding some sort of gnu-config-initial-native recipe would
fix the latter problem, but this would be ugly because it would need special-casing
in (at least) autotools.bbclass, and in any case this still wouldn't solve
the problem of autoconf itself depending on g-c-n.
So, the easiest solution to the problem at hand is to arrange for those
few recipes that depend on g-c-n but not autoconf-native to gain that
latter dependency as well.
Signed-off-by: Phil Blundell <pb@pbcl.net> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Phil Blundell [Fri, 10 Jan 2014 12:57:02 +0000 (12:57 +0000)]
kernel-module-split: Remove extraneous call to depmod from module postinst
During rootfs construction, image.bbclass will call depmod after all the
modules are installed. There's no need to run it from the postinst when
operating in offline root mode.
Signed-off-by: Phil Blundell <pb@pbcl.net> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Phil Blundell [Thu, 9 Jan 2014 11:16:10 +0000 (11:16 +0000)]
libsoup: Remove libproxy from DEPENDS
Although libsoup did use to support direct usage of libproxy, it hasn't
done so for some time. Worse, if libsoup depends on libproxy then it
is impossible to build libproxy against webkit since webkit itself
depends on libsoup in some configurations. Fix this by removing the
extraneous entry from DEPENDS.
Signed-off-by: Phil Blundell <pb@pbcl.net> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Phil Blundell [Thu, 9 Jan 2014 11:18:26 +0000 (11:18 +0000)]
qemu: Remove stray .git files after unpacking
The qemu tarball ships some extraneous .git metadata in the dtc/ and pixman/
subfolders, containing a reference to the path "../.git/modules" which
doesn't exist. The presence of these files will confuse git if it happens
to recurse into the qemu source folder during an operation on some parent
directory, for example "git clean" at a higher level. Avoid that problem
(mostly) by removing them immediately after the sources are unpacked.
Signed-off-by: Phil Blundell <pb@pbcl.net> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
libpam: add a new 'nullok_secure' option support to pam_unix
'null_ok_secure' option was fixed. Since that commit 'su' stopped working in
pseudo terminals (created in X environments) for root-accounts with empty
password.
Background: The PAM configuration for 'su' includes 'common-auth' which uses
'nullok_secure' option for pam_unix.
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Chen Qi [Mon, 23 Dec 2013 08:16:47 +0000 (03:16 -0500)]
busybox: enable SEAMLESS_XZ by default
We have enabled SEAMLESS_GZ and SEAMLESS_BZ2 in defconfig, it's
reasonable for us to enable SEAMLESS_XZ as well. Otherwise, we
couldn't extract tar.xz file while we could extract tar.gz file.
Such situation would be somewhat confusing.
Cristian Iorga [Wed, 8 Jan 2014 10:13:26 +0000 (12:13 +0200)]
qemu: upgrade to 1.7.0
linux-user-Handle-SOCK_CLOEXEC-NONBLOCK-if-unavailab.patch
file no longer needed, included in upstream.
qemu-native tested on all architectures,
host machine is Ubuntu Linux 13.10 x86-64.
Basic X11 and networking tests performed.
Paul Barker [Mon, 6 Jan 2014 11:55:14 +0000 (11:55 +0000)]
libarchive: Upgrade to v3.1.2
All patches against libarchive in oe-core appear to be merged into the latest
release. The license checksum has changed because a couple of referenced files
have been renamed but there is no change to the license terms themselves.
Signed-off-by: Paul Barker <paul@paulbarker.me.uk> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Ming Liu [Thu, 14 Nov 2013 10:51:28 +0000 (18:51 +0800)]
sstate: Add optimizing logic for crosssdk setscene dependencies
This patch mainly aims to add optimisation for crosssdk setscene dependency
validating which we haven't handled in current logic, and which I think we
could have as we've already implemented to native/cross, although there
are albeit not many crossdk tasks, we could still get some performance
enhancement.
Signed-off-by: Ming Liu <ming.liu@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Kevin Baker [Sat, 4 Jan 2014 04:46:05 +0000 (22:46 -0600)]
screen: update debian patchset to version 4.0.3-14
This recipe used an old version of the debian update patchset from Lenny which
does not exist anymore on Debian's FTP servers. Update to the latest set
version 4.0.3-14 which also includes some minor bug fixes and security updates.
Cleanly builds and runs in our armv7a-vfp-neon environment.
Signed-off-by: Kevin Baker <kbaker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Martin Jansa [Sat, 4 Jan 2014 17:20:20 +0000 (18:20 +0100)]
python: explicitly disable bluetooth.h check
* bluetooth.h is autodetected from sysroot and influences 2 python files:
/usr/include/python2.7/pyconfig-32.h
/usr/lib/python2.7/lib-dynload/_socket.so
* it doesn't link with bluez, so it wasn't detected by
test-dependencies.sh, but still causes undeterministic builds and
should be fixed
* we can use PACKAGECONFIG, but I don't expect many people to use bt
support in python-socket
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Koen Kooi [Mon, 6 Jan 2014 11:01:25 +0000 (12:01 +0100)]
grub: add git version
Recently grub git gained support for ARM builds (using u-boot or EFI as first stage loader) and with 2 extra patches we get support for 64-bit ARM as well.
Buildtested for genericarmv7a, genericarmv8 and qemux86. The genericarmv8 build fails in do_package/strip due to a binutils problem.
Signed-off-by: Koen Kooi <koen.kooi@linaro.org> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Chen Qi [Mon, 6 Jan 2014 07:27:34 +0000 (15:27 +0800)]
initscripts: split the functions script into a separate package
Many SysV init scripts need the /etc/init.d/functions script. But
this script is part of the initscripts package. As a result, the
initscripts package should always be installed into the system to
avoid errors when starting daemons. However, it makes no sense to
install the initscripts package into a systemd based image, because
what the init scripts provide has already been provided by the systemd.
On the other hand, the functions script might be still needed in a
systemd based image because other init scripts such as distcc might need
it.
The above situation leads to a natural separation of the functions script
from the initscripts package. And this patch does so. It separates the
functions script into initscripts-functions packages.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
*Some* packages have migrated away from gconf to gsettings. Not all
packages have done this, so anything still using gconf is now broken.
Note that almost nothing in oe-core uses gsettings as we're still
using the GTK+ 2 stack generally.
The problem is that people are using gnome.bbclass which is a kitchen
sink class, inheriting many other classes that may or may not be used.
To resolve this warning gnome.bbclass should extend the sanity check
whitelist with --disable-schemas-install as it is the class causing
gconf.bbclass to be inherited on recipes that don't use gconf.
Philip Balister [Sun, 29 Dec 2013 18:27:38 +0000 (13:27 -0500)]
db : Include C++ binding library in Berkeley DB recipe.
Create the C++ library for db. This library is packaged in a new
package so the db package only contains the c library. This prevents
existing users from adding libstdc++ to the package DEPENDS.
Signed-off-by: Philip Balister <philip@balister.org> Signed-off-by: Douglas Geiger <doug.geiger@bioradiation.net> Signed-off-by: Steve Arnold <esteve@gentoogeek.org> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Chong Lu [Fri, 3 Jan 2014 06:51:30 +0000 (14:51 +0800)]
quilt: enable ptest support
Install quilt test suite and run it as ptest.
Exclude delete.test and mail.test.
delete.test need execute with normal user and mail.test need depend
on mail agent, else these test cases will fail. So exclude them order
to make ptest all pass.
Signed-off-by: Chong Lu <Chong.Lu@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Mike Crowe [Fri, 3 Jan 2014 14:28:39 +0000 (14:28 +0000)]
valgrind: rrecommend libc6-dbg on all architectures
It seems that valgrind requires debug symbols to be available on all
architectures so there's no reason to treat arm and powerpc specially.
Ensuring that libc6-dbg is installed avoids errors like this on x86:
valgrind: Fatal error at startup: a function redirection
valgrind: which is mandatory for this platform-tool combination
valgrind: cannot be set up. Details of the redirection are:
valgrind:
valgrind: A must-be-redirected function
valgrind: whose name matches the pattern: strlen
valgrind: in an object with soname matching: ld-linux.so.2
valgrind: was not found whilst processing
valgrind: symbols from the object with soname: ld-linux.so.2
valgrind:
valgrind: Possible fixes: (1, short term): install glibc's debuginfo
valgrind: package on this machine. (2, longer term): ask the packagers
valgrind: for your Linux distribution to please in future ship a non-
valgrind: stripped ld.so (or whatever the dynamic linker .so is called)
valgrind: that exports the above-named function using the standard
valgrind: calling conventions for this platform. The package you need
valgrind: to install for fix (1) is called
valgrind:
valgrind: On Debian, Ubuntu: libc6-dbg
valgrind: On SuSE, openSuSE, Fedora, RHEL: glibc-debuginfo
valgrind:
valgrind: Cannot continue -- exiting now. Sorry.
Signed-off-by: Mike Crowe <mac@mcrowe.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Petter Mabäcker [Fri, 3 Jan 2014 18:11:43 +0000 (19:11 +0100)]
cdrtools-native: remove unused FILESPATH
Usage of FILESPATH is discouraged, since it can make recipes harder to
bbappend. Instead FILESEXTRAPATHS should be used to extend the path.
However in cdrtools no FILESPATH additions are currently needed so
instead it should be removed.
Robert Yang [Fri, 27 Dec 2013 13:46:10 +0000 (21:46 +0800)]
make: upgrade to 4.0
Upgrade to 4.0, remove the following patches since they have been fixed
by the upgrade:
- expand_MAKEFLAGS.patch
- intermediate-target-bugfix.patch
- make-savannah-bug30612-fix_white_space.patch
- make-savannah-bug30612-handling_of_archives.patch
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Robert Yang [Thu, 2 Jan 2014 08:37:34 +0000 (03:37 -0500)]
logrotate: two minor fixes
* Change EXTRA_OEMAKE = "CC='${CC}'" to "", the root cause is the "-e"
from EXTRA_OEMAKE makes the CFLAGS from the env overrides the one
defined in logrotate/Makefile which causes build failures, so we need to
set EXTRA_OEMAKE to "", set it to CC='${CC}' would confuse the user
since CC='${CC}' has been defined in the env.
* Merge do_install() and do_install_append()
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Robert Yang [Thu, 2 Jan 2014 08:29:45 +0000 (03:29 -0500)]
logrotate: upgrade to 3.8.7
* Upgrade to 3.8.7
* Rename the patches dir from logrotate-3.8.1 -> logrotate
* Remove grotate-CVE-2011-1548.patch since it had been fixed
* Update act-as-mv-when-rotate.patch and update-the-manual.patch to make
them work with the higher version, and send them to the upstream
* Fix the HOMEPAGE
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Robert Yang [Tue, 31 Dec 2013 17:25:17 +0000 (01:25 +0800)]
e2fsprogs: upgrade to 1.42.9
* Upgrade to 1.42.9
* Remove the following patches since they have been merged/fixed by
upstream:
- debugfs-extent-header.patch
- debugfs-sparse-copy.patch
- debugfs-too-short.patch
- e2fsprogs-fix-tests-f_extent_oobounds.patch
- fallocate.patch
* The populate-extfs.sh had been merged by the upstream, but I'd like to
go on using the previous one which is from our meta layer, they are a
little different, and the script would be dropped when we use the mke2fs
to populate the rootfs.
* Sumitted the patch for populate-extfs.sh (from Søren Holm) to upstream.
* Submitted fix-icache.patch to upstream, I wrongly thought it was not
applicable to the upstream, but it does.
* Join the do_install() and do_install_append() together.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Robert Yang [Wed, 1 Jan 2014 03:27:09 +0000 (11:27 +0800)]
mime.bbclass: remove the --disable-update-mimedb
Gnome and other freedesktop's mime data has been moved to
shared-mime-info, and there is no such an option any more, we already
have this recipe, so remove it, this will fix the warning:
configure was passed unrecognised options: --disable-update-mimedb
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
