Eric Bénard [Sat, 10 Sep 2011 09:17:57 +0000 (11:17 +0200)]
qt4: fix generated sdk
- qt4-tools-nativesdk : actually the qmake binary which gets installed
comes from the native recipe. This patch fix this problem by launching
configure twice : once to compile qmake using the right toolchain for
nativesdk, and a second time using the native qmake to compile all the
other tools for the nativesdk. Then we install the right qmake.
- mkspec : the link actually created in qt4-tools-nativesdk's
do_install point to nowhere so remove it and generate the link in
meta-toolchain-qte as it's the only place where we have all the variable
to create it.
- toolchain_create_sdk_env_script_append : we need to add OE_QMAKE_CFLAGS,
OE_QMAKE_CXXFLAGS and OE_QMAKE_LDFLAGS else the sdk won't find these
variables that are inserted by qmake in the Makefiles.
- with this patch, oe-core generates a working meta-toolchain-qte which
can compile a small example and is properly recognized by qtcreator (this
brings oe-core's meta-toolchain-qte to oe-dev's functional state).
Richard Purdie [Thu, 29 Sep 2011 22:48:20 +0000 (23:48 +0100)]
bitbake.conf: Start using parallel make for do_install
Most piece of software now support parallel make install. Enable
this by default using the value of PARALLEL_MAKE. In a similar way
to PARALLEL_MAKE we can disable this for broken recipes with:
PARALLEL_MAKEINST = ""
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 29 Sep 2011 22:30:26 +0000 (23:30 +0100)]
base.bbclass: Drop unneeded dependency
patch depends on unpack
configure depends on patch
We simply don't need a configure dependency on unpack. This simplifies
the dependencies of every recipe slightly and should make bitbake
slightly faster at resovling dependency graphs.
It also makes the .dot dependency graphs slightly more readable by
removing noise.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 29 Sep 2011 22:29:08 +0000 (23:29 +0100)]
autotools: Disable dependency tracking
There is no point in us spending time generating a ton of dependency
information in source directories when we don't actually use it
for our one off builds. We might as well disable the option and
take the speed up.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 29 Sep 2011 22:26:07 +0000 (23:26 +0100)]
Remove help2man dependency
The help2man script is pretty useless to us. It requires to run the target
binary to extract help information which is not possible for any of our
cross compiled target binaries.
We're not interested in man pages for -cross/-native tools.
It therefore makes no sense to have this as a core build dependency.
This patch removes the dependeny and replaces it with a script
returning false. This will trigger autotool's missing utility
to use the copy of the man page included with the sources which
is what would already happen when we tried to run cross compiled
binaries anyway.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This core does not have altivec, so we disable it in the build,
also reestablish the config option to enable/disable building
with altivec
If SPE is not detected we always build with altivec which is wrong. This
will check to make sure altivec is enabled and pass build options
through accordingly
Signed-off-by: Matthew McClintock <msm@freescale.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add proper deps for nfs-utils, util-linux, and strace
These packages need these deps for the RPM generation stage:
error: Failed dependencies:
| /usr/bin/python is needed by nfs-utils-1.2.3-r2.ppce5500
| /usr/bin/perl is needed by util-linux-2.19.1-r4.ppce5500
| /usr/bin/perl is needed by strace-4.5.20-r2.ppce5500
Signed-off-by: Matthew McClintock <msm@freescale.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add support for generating ext2.bz2{, .u-boot} images
Machines can now add ext2.bz2 and ext2.bz2.u-boot images as a
generated image type. This also adds an extra parameter to
oe_mkimage which is the image compression type for mkimage
Signed-off-by: Matthew McClintock <msm@freescale.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 29 Sep 2011 22:35:16 +0000 (23:35 +0100)]
insane: Promote most warnings into QA errors by default
We agreed a while back that we'd start ERRORing on QA issues, not just
WARNING about them. This patch changes the default QA levels to error
on everything except ldflag and useless rpath issues.
It also stops giving out QA warnings about desktop files since it
adds noise to the build and until someone plans to seriously tackle
that problem, its pointless.
We can promote the ldflag and rpath warnings too, when the issues in
OE-Core are resolved (they're mostly there now, just a few more to go).
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andrew Gabbasov [Wed, 28 Sep 2011 12:07:58 +0000 (16:07 +0400)]
apmd: use ${HOST_SYS}-libtool
libtool-cross uses ${HOST_SYS}- prefix while building and installing.
In some cases that may be different from ${TARGET_PREFIX}, that is currently
used in apmd recipe. It's better to have them consistent.
Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andrew Gabbasov [Wed, 28 Sep 2011 12:09:52 +0000 (16:09 +0400)]
apr: use ${HOST_SYS}-libtool
libtool-cross uses ${HOST_SYS}- prefix while building and installing.
In some cases that may be different from ${TARGET_PREFIX}, that is currently
used in apr recipe. It's better to have them consistent.
Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andrew Gabbasov [Wed, 28 Sep 2011 12:10:30 +0000 (16:10 +0400)]
libpam: add flex-native to DEPENDS
flex-native is required for building libpam. Although this dependency
is now fulfilled indirectly through bison recipe, having an explicit one
would be preferable.
Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andrew Gabbasov [Wed, 28 Sep 2011 12:10:12 +0000 (16:10 +0400)]
cmake.bbclass: adding extendable extra variable to root path
Introducing a hook variable for various .conf files to add
extra path elements to cmake. For example, it can be used
in external toolchain conf file to add ${TOOLCHAIN_PATH}
to cmake root path search.
Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Wed, 28 Sep 2011 15:28:25 +0000 (16:28 +0100)]
scripts: use OE_TMPDIR instead of TMPDIR external variable
On OpenSUSE within an X session, TMPDIR is set to the system temporary
directory (/tmp) which is incorrect for these scripts. Thus, change
runqemu and oe-setup-rpmrepo to use OE_TMPDIR from the external
environment rather than TMPDIR.
Fixes [YOCTO #1530]
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
currently we have allarch type of recipes, which may still have
architecture dependency, like x11-common. So we need to drop the
handling to allarch in multilib case.
Also remove the PV postfix in python-pygobject DEPENDS, since multilib
code will treat a native package multilib capable.
ghostscript has it's own hacky check for time.h which hard-codes paths, this
means in the native case it fails on systems such as Ubuntu 11.10 where the
location of time.h has changed. Further it means the target build has had a
host-intrusion issue.
This patch disables the check for time.h, future releases of ghostscript
use standard autotools checks for time.h's location.
rpm-native was reading from /usr/share/misc/magic which is wrong
it needs to be set to read from the sysroot. This also adds wrappers
to the rpm-build tools to ensure they know were to find the macros that
point to the right directories.
Khem Raj [Tue, 16 Aug 2011 22:20:15 +0000 (15:20 -0700)]
gcc-4.6: Use --with-linker-hash-style configure option
Depending upon what hash style is in use this uses
the right flag for setting the hash style type. This
fixes the QA errors about missing GNU hash style reported
in gcc-runtime build particularly libgcc
Khem Raj [Tue, 16 Aug 2011 22:17:33 +0000 (15:17 -0700)]
conf,recipes: Add new variable LINKER_HASH_STYLE
LINKER_HASH_STYLE in OE is set to either 'sysv' or 'gnu'
depending upon processor architecture e.g. mips does not support
gnu hash style so is uses sysv
besides 'sysv' and 'gnu' third option is to set it to 'both' we do
not do that by default but user can still set it
Simon Busch [Sun, 25 Sep 2011 12:57:44 +0000 (14:57 +0200)]
qt4-embedded: move qte.sh profile script into own package
We don't want to pull the whole qt suite in our images when only installing a tiny setup
which needs the profile script for several configuration issues. This moves the qte.sh
profile script into a qt4-embedded-conf package which is automatically installed when an
image depends on qt4-embedded.
Simon Busch [Sun, 25 Sep 2011 12:57:43 +0000 (14:57 +0200)]
meta: qt4: fix postprocessing of pkg-config files
When building qt4-embedded the generated and cleaned pkg-config files for qt are wrong.
The Cflags variable contains something like ${includedir}/qtopia/QtCore where
${includedir} is already /usr/include/qtopia/QtCore.
This patch reverts the fix up of the Cflags variable implemented in do_install.
Richard Purdie [Mon, 26 Sep 2011 20:12:25 +0000 (21:12 +0100)]
pseudo: Fix QA warnings
This fixes two QA warnings:
a) Debug files being contained in the main package (by adding
an appropriate FILES expression)
b) Stop hardcoding the RPATH in the nativesdk case since our
path is on the loaders default search path
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 26 Sep 2011 20:10:45 +0000 (21:10 +0100)]
sqlite3: Fix nativesdk packaging/QA warnings
When building sqlite3-nativesdk, there was a warning about debug files in the
main package. This was due to the order of items in PACKAGES with -dbg after
the main package which breaks an assumption native.bbclass makes. Changing
the order fixes the packaging problem with no change to the normal target
packaging.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 26 Sep 2011 20:11:29 +0000 (21:11 +0100)]
grub: Fix insane/QA architecture warning
There is QA warning about this package for an architecture mismatch but
this is inappropriate in this case since the bootloader needs 32 and 64
bit code. We therefore flag the QA check to be skipped.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 26 Sep 2011 20:13:21 +0000 (21:13 +0100)]
gcc: Fix two QA issues
a) There is a QA warning from a .so being present in a main package.
In the case of the plugin library for gcc, this is allowed.
b) Remove the unwanted libiberty.a file with the strange path. We
don't need/want this and this removes an unpackaged file warning.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Mon, 26 Sep 2011 15:55:56 +0000 (11:55 -0400)]
linux-yocto/meta: eg20t and live boot config changes
Merging the following configuration changes:
67a46a6 meta/common-pc-64: enable live booting for common-pc-64 1010905 meta/common-pc: enable live booting for common-pc b3c5fa7 meta/atom-pc: enable live booting for atom-pc 41c090e meta: update boot live config and move it to cfg/ d51b0e7 eg20t: update config options
The first 4 make the live-boot configuration shared and then reuse
them for the boards that currently are live bootable. The eg20t
is a cleanup of obselete kernel options and is part of the cleanup
of options for the 3.0 kernel.
[YOCTO: #940]
[YOCTO: #686]
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com> Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
During the avahi-ui split the RDEPENDS line got dropped leading to: http://git.angstrom-distribution.org/cgi-bin/cgit.cgi/testlab/commit/?h=yocto&id=ae0db0eb51bde42f09b9513842c3b745c37c60d5
As per gcc PR 11147, libffi installs headers into a target dependent
place (/usr/lib/....). Include a rule to include those files into
libffi-dev package.
gcc: fix possible problems with nscd compilation during eglibc-nativesdk build
Long time ago a066e7ca90a28d5681c5fa895a29e999ed7c88b was committed to
address possible problems with compilation of nscd during
eglibc-nativesdk build. Problems were related to the way gcc searches
for headers to check if it should enable it's own stack smash protection
bits or it can relay on eglibc for it.
However after 934d38530c9a67562e53d4034aee5531f0f26750 things got
broken, as for gcc-crosssdk-intermediate packages:
1) EXTRA_OECONF is ignored
2) headers are installed in a different location than expected by that
patch.
This results in eglibc-nativesdk build broken on some systems (e.g. mine
Debian x86_64 squeeze). Fix that by providing with-headers options to
crosssdk-intermediate gcc configuration.
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
icecc-create-env-native: provide the script right in the tree
There is no point in downloading a tarball with no clear upstream (other than
icecc itself) and then patching it. Rather put new script in the source tree.
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
bitbake.conf: change ccache path to use MULTIMACH_HOST_SYS
Currently if I build packages for several targets (e.g. for armv5te tosa
and for armv7a beagleboard) oe will use single ccache dir for both of
those targets: build/ccache/arm-oe-linux-gnueabi. However those targets
use different opcodes, different features and binaries created for one
of those targets wont't run on the lower one. So use MULTIMACH_HOST_SYS
for ccache dir, so that it uses something like
build/ccache/armv5te-oe-linux-gnueabi dir.
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
gnome-doc-utils: Prepend PKG_CONFIG_SYSROOT_DIR to the path returned from PKG_CONFIG
If we build say gnome based image on a build system which does not have
gnome e.g. kubuntu then packages like gedit do not build since it uses
gnome files from host system which are non existent on kubuntu
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 23 Sep 2011 14:02:58 +0000 (15:02 +0100)]
base.bbclass: Drop legacy code that is no longer useful/used
The code displaying console events has been handled by the bitbake
UI since 1.8 so the legacy code path can be removed. If a log event
is wanted, there are much better (and more complete) ways to do this
so remove the EVENTLOG code too.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 23 Sep 2011 07:57:31 +0000 (08:57 +0100)]
autotools.bbclass: Set the dynamic linker search path for libtool correctly
libtool obtains the search path from /etc/ld.so.conf and hardcodes /usr/lib
and /lib. This results in host contamination and variable sets of RPATH
values ending up in binaries.
By exporting the correct values for all autotools recipes we avoid this.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
meta-oe has a version of tcp-wrappers 7.6 recipe
which has virtually no differences with the one from
oe-core. So with this patch we can remove the recipe
from meta-oe
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Added a native sstate post install function to fix the links
created between /lib and /usr/lib for the library files. These
links could point to an invalid build area when using shared state.
Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
