Richard Purdie [Sat, 25 Aug 2012 15:46:12 +0000 (16:46 +0100)]
e2fsprogs: Drop fsck binary
There are conflicts in the native sysroot over the fsck binary. Whilst
providing people with alternatives is good to a point, the source in util-linux
is clearly the copy which is now being maintained and moved forwards which the copy
in e2fsprogs being older and with less features.
The simplest solution is simply to no longer offer this and used the maintained
version from util-linux.
Richard Purdie [Sat, 25 Aug 2012 15:45:33 +0000 (16:45 +0100)]
e2fsprogs: Fully remove/disable blkid
We've removed libblkid from e2fsprogs as it was being provided by util-linux. The blkid
binary is also provided there and providing it within e2fsprogs too, linked against
a different and potentially incompatible library we're likely asking for trouble.
It also leads to inconsistencies in the native sysroot since we don't consistently
get one given provider of blkid (but always the liblibid from util-linux).
To clean this up, fully remove/disable the blkid binary and package.
Roy.Li [Wed, 8 Aug 2012 01:44:26 +0000 (09:44 +0800)]
bitbake: compile tar-replacement firstly
Compiling tar-replacement or not is decided by version of host tar,
if the host tar version is lower than 1.23, Compiling tar-replacement
is needed.
When doing popoluate tar-replacement sysroot to write the tar to
sysroot, but writing is not finished. other packages probably
use the being written tar to unzip file, which will lead to failure
and report the below error:
"bitbake_build/tmp/sysroots/x86_64-linux/usr/bin/tar: Text file busy"
Now we compile tar-replacement firstly to ensure that a being written
tar command will not be used.
Tom Zanussi [Tue, 7 Aug 2012 21:11:14 +0000 (16:11 -0500)]
yocto-bsp: generate default properties even if json specified
Users seem to want to specify incomplete property sets when using json
input. Allow this by generating default properties before the
user-specified properties are applied; the user will then get the
defaults for any unspecified values, and avoid cryptic backtraces.
Tom Zanussi [Mon, 6 Aug 2012 05:56:55 +0000 (00:56 -0500)]
yocto-bsp: use rstrip() for assignment lines
strip() isn't necessary and causes unintended formatting changes in
the output; rstrip() remove the trailing newlines as intended while
leaving indenting whitespace intact.
Jason Wessel [Wed, 25 Jul 2012 17:47:28 +0000 (12:47 -0500)]
terminal.py: use unique ids for screen sessions
When running multiple sets of builds on the same
system, it is hard to distinguish which build belongs
to which screen session and you can end up resuming
the wrong session.
The simple solution is to just append the process
id to the screen session invocation to make each
unique.
Bruce Ashfield [Wed, 22 Aug 2012 04:42:05 +0000 (00:42 -0400)]
kernel-yocto: don't require meta branch for custom repos
The commit a9f11849 [linux-yocto: explicitly export KMETA to scripts]
allows the meta branch name to be changed by exporting it to all
phases of the build.
But if a custom kernel without a meta branch is built, we end up
passing an empty string to the creation scripts, which breaks the
build since input is expected.
Inhibiting the export of KMETA to the creation scripts when empty
fixes the problem.
Bruce Ashfield [Wed, 22 Aug 2012 05:18:05 +0000 (01:18 -0400)]
kernel-yocto: set master branch to a defined SRCREV
To support custom repositories that set a SRCREV and that only have
a single master branch, do_validate_branches needs a special case
for 'master'. We can't delete and recreate the branch, since you
cannot delete the current branch, instead we must reset the branch
to the proper SRCREV.
Liang Li [Wed, 22 Aug 2012 08:02:33 +0000 (16:02 +0800)]
kernel.bbclass: fix INC directory for SLANG
In the kernel's tools/perf/Makefile CFLAGS was previously hard coded to
contain "-I/usr/include/slang" to work with hosts that have
"/usr/include/slang/slang.h" as well as hosts that have
"/usr/include/slang.h". This path can cause compile warnings like:
"ERROR: Nothing PROVIDES 'grub-efi-x86_64-native'" is returned when
trying to build meta-intel/meta-cedartrail with DEFAULTTUNE = "core2-64"
Used TRANSLATED_TARGET_ARCH instead of TARGET_ARCH in grub-efi.bbclass
kmod-native_git.bb: fix builds for hosts with older libc
kmod will fail to build with the following error because O_CLOEXEC is
not defined:
| libkmod/libkmod-module.c: In function 'kmod_module_get_initstate':
| libkmod/libkmod-module.c:1640: error: 'O_CLOEXEC' undeclared (first use in this function)
| libkmod/libkmod-module.c:1640: error: (Each undeclared identifier is reported only once
| libkmod/libkmod-module.c:1640: error: for each function it appears in.)
| libkmod/libkmod-module.c: In function 'kmod_module_get_refcnt':
| libkmod/libkmod-module.c:1754: error: 'O_CLOEXEC' undeclared (first use in this function)
| libkmod/libkmod-module.c: In function 'kmod_module_get_sections':
| libkmod/libkmod-module.c:1913: error: 'O_CLOEXEC' undeclared (first use in this function)
| libkmod/libkmod-file.c: In function 'kmod_file_open':
| libkmod/libkmod-file.c:282: error: 'O_CLOEXEC' undeclared (first use in this function)
| libkmod/libkmod-file.c:282: error: (Each undeclared identifier is reported only once
| libkmod/libkmod-file.c:282: error: for each function it appears in.)
Since we are only using kmod-native for depmod, and it's a non-threaded
user of this libary being built this should be safe to override O_CLOEXEC.
Keep in mind this is ONLY effecting the native builds and not what is
being shipped in the root file system.
Jackie Huang [Fri, 24 Aug 2012 13:29:06 +0000 (21:29 +0800)]
gdk-pixbuf: call gtk-update-icon-cache with -t in postinstall script
gtk-update-icon-cache is called for each subdir of ${datadir}/icons,
but there are some themes without an index.theme file such as
xcursor-transparent, so add -t(--ignore-theme-index) option to avoid
errors of gtk-update-icon-cache.
Andy Ross [Tue, 14 Aug 2012 22:44:22 +0000 (15:44 -0700)]
task-core-lsb: Don't pull in eglibc-pic via RDEPENDS
The task-core-lsb-runtime-add subpackage includes eglibc-pic as an
RDEPEND. That's incorrect, eglibc-pic is nothing but 22MB of static
libraries and should never appear as a runtime dependency.
Saul Wold [Thu, 23 Aug 2012 18:33:18 +0000 (11:33 -0700)]
iputils: chmod in pkg_postinst are redundant
With pseudo the cmod in pkg_postinst is actaully redundant since
the do_install uses install -m 4555 and pseudo is able to track that
between the install time and the rootfs package installation, so the
perms are correct.
Richard Purdie [Fri, 24 Aug 2012 16:12:39 +0000 (16:12 +0000)]
distutils/steuptools: Fix files layout and unbreak builds
The last two distutils changes progressivly broke the builds. Firstly they
moved things from the site_packages directory to being higher up the tree
which introduced package QA warnings as a side effect. Secondly, it interacts
badly with setuptools which passes in --root=${D} itself.
This patch restores the original directory layout, hence fixing the QA
warnings and also passes extra options to setuptools to deal with the
--root option it passes.
Andy Ross [Thu, 23 Aug 2012 17:32:31 +0000 (17:32 +0000)]
ld: -rpath must search under sysroot
The -rpath argument would search the host filesystem for libraries,
even when a sysroot was defined. For cross toolchains with targets
compatible with the host architecture this can find incorrect
libraries. Leave -rpath-link unmodified, as build systems in the wild
are already using this to point to host directories.
Paul Eggleton [Mon, 20 Aug 2012 16:05:47 +0000 (17:05 +0100)]
poky.conf: restore Ubuntu 12.04 LTS to tested distro list
Apparently some people don't apply the normal updates to their Ubuntu
installations, in which case they will not have the version updated to
12.04.1. Since the distinction between the two is fairly minimal,
restore the old version string in addition to the new one.
Paul Eggleton [Mon, 20 Aug 2012 13:20:41 +0000 (14:20 +0100)]
libiconv: skip recipe if using eglibc
libiconv is provided for use with uClibc - if you build it together with
eglibc (which already PROVIDES virtual/libiconv) you can end up with
dependency problems during do_rootfs.
Radu Moisan [Wed, 22 Aug 2012 09:22:22 +0000 (12:22 +0300)]
busybox: Upgrade to upstream 1.20.2
Updated defconfig file from generated .config
Disabled CONFIG_LSOF,CONFIG_CROSS_COMPILER_PREFIX, and CONFIG_EXTRA_CFLAGS
for backwards compatibility.
Added busybox-1.20.2-kernel_ver.patch from upstream.
Radu Moisan [Tue, 21 Aug 2012 11:48:38 +0000 (14:48 +0300)]
grub-efi-native: Update to version 2.00
Added -Wno-unused-result because of a read() with no result, and causing
error due to -Werror
Updated grub recipe because of a patch name change
Removed unnecessary patches (merged upstream).
Paul Eggleton [Tue, 21 Aug 2012 13:17:10 +0000 (14:17 +0100)]
scripts/combo-layer: specify branch when getting current revision
Handle the case during update where the configured branch is not
currently checked out in the component repository by just specifying
it in the places where it was not previously.
Xin Ouyang [Mon, 20 Aug 2012 19:36:46 +0000 (14:36 -0500)]
grub-1.99: Add flex-native to DEPENDS
[YOCTO #2916]
On some systems the host version of flex may be too old to work properly,
instead we depend on flex-native to ensure that we have the proper version
available to us. (Flex version 2.5.35 or newer is requried.)
Khem Raj [Wed, 22 Aug 2012 17:59:48 +0000 (10:59 -0700)]
linux-yocto: Remove parens from COMPATIBLE_MACHINE regexp
This helps in appending to this regexp from bbappends
coming from other layers who want to leverage linux-yocto
with minimum tweaks by using a standard prefefined machine
from linux-yocto
Khem Raj [Wed, 22 Aug 2012 17:59:47 +0000 (10:59 -0700)]
runqemu: Honor KERNEL and ROOTFS settings from environment
Expand the usage to be able to specify KERNEL and ROOTFS
on commandline. This helps in using the script for booting
images that are essentially not part of OE-Core
Richard Purdie [Wed, 22 Aug 2012 12:53:11 +0000 (13:53 +0100)]
bitbake: methodpool: Clean up the parsed module list handling to be slightly less insane
This removes some dubious functions and replaces them with a simpler, cleaner
API which better describes what the code is doing. Unused code/variables are
removed and comments tweaked.
Richard Purdie [Wed, 22 Aug 2012 19:33:39 +0000 (20:33 +0100)]
bitbake: cooker: Ensure parsing failures stop the build
Currently parsing failures still allow bitbake to continue on and try
and execute a build. This is clearly a bad idea and this patch adds in
more correct error handling and stops the build.
The use of sys.exit is nasty but this patches other usage in this function
so is at least consisent and its better than the current situation of
trying to execure a half parsed set of recipes. There are probably better
ways this could be improved to use to stop the build.
Richard Purdie [Wed, 22 Aug 2012 19:02:39 +0000 (20:02 +0100)]
bitbake: utils.py: Once we've printed errors messages, suppress the backtrace
Adding a generic backtrace to better_compile and better_exec is pointless,
therefore reraise the exception as a bb.BBHandledException so the
generic code doesn't confuse the user even more.
Richard Purdie [Wed, 22 Aug 2012 19:01:12 +0000 (20:01 +0100)]
bitbake: utils.py: Try harder to extract good traceback information by querying the datastore
Currently as soon as execution passes outside the code fragment being
executed by better_exec, we don't get any good traceback information,
just a likely obscure reference to some function name which may
or may not be identifiable.
This patch adds code to query the datastore if present, allowing a more
meaningful back trace to be displayed in many cases.
Paul Eggleton [Wed, 22 Aug 2012 15:14:38 +0000 (16:14 +0100)]
bitbake: cooker: fix traceback when using -b with skipped recipe
If a recipe is skipped during parsing for whatever reason, check and
report this as an error rather than trying to use the data that is sent
back and failing.
Paul Gortmaker [Fri, 17 Aug 2012 15:24:19 +0000 (11:24 -0400)]
routerstationpro: move board off 3.0 and onto the 3.4 kernel
The updated patch series to support this target is in place
on the BSP specific branch in the 3.4 kernel tree now[1], so
we can move it ahead off of the old 3.0 kernel.
Andy Ross [Fri, 17 Aug 2012 15:53:11 +0000 (08:53 -0700)]
libtool: normalize link paths before considering for RPATH
Libtool may be passed link paths of the form "/usr/lib/../lib", which fool
its detection code into thinking it should be included as an RPATH in
the generated binary. Normalize before comparision.
distutils.bblass: change order of args to install step
This let's the user override install-lib argument again if it needs
to be something else, otherwise things like python-setuptools
won't be able to modify the install-lib dir
This fixes a new issue exposed by my previous distutils patch
that fixed the python modules default install location. Also,
it removes running the install step twice which was inadvertant
Darren Hart [Tue, 21 Aug 2012 16:27:16 +0000 (09:27 -0700)]
README: Clarify where to send patches
There is still a great deal of confusion over where patches for the
various components of the poky repository should be sent. Attempt to
mitigate this by including a blurb in the poky README similar to what
other layers and repositories do.
V2: Call out the special care required when dealing with the scripts
directory per Saul Wold's recommendation.
V3: Recommend applying patches to oe-core prior to sending patches to
the oe-core mailing list.
Signed-off-by: Darren Hart <dvhart@linux.intel.com> CC: Richard Purdie <richard.purdie@intel.com> CC: Paul Gortmaker <paul.gortmaker@windriver.com> CC: Paul Eggleton <paul.eggleton@linux.intel.com> CC: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Tue, 21 Aug 2012 17:08:29 +0000 (18:08 +0100)]
conf/bitbake.conf: allow cache reuse between bitbake, bitbake-layers and hob
BB_ENV_EXTRAWHITE is set by the bitbake wrapper script but not by
bitbake-layers, and thus it was affecting the data hash and thus
invalidating the cache. Any variables passed through from the
environment by being added to this variable should themselves
influence the hash already if they are not already excluded from,
the data hash, so BB_ENV_EXTRAWHITE can be excluded.
Similarly, DISABLE_SANITY_CHECKS is set by the hob script in order to
have sanity checks run later, but not set by the bitbake wrapper script.
It doesn't actually affect the build itself and so its value can also
be excluded from the data hash.
These changes prevent unnecessary reparsing between executions of
bitbake, bitbake-layers, and the initial pseudo build within the hob
script.
Signed-off-by: Konrad Scherer <Konrad.Scherer@windriver.com> Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Reword commit message, rebase to latest bitbake.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Tue, 21 Aug 2012 17:11:47 +0000 (18:11 +0100)]
bitbake: bitbake-layers: use parsable title for show-appends
Print title text surrounded by === as we do with other subcommands so
that you can filter it out in scripts. (You will also want to filter out
lines starting with "WARNING:" if you just want the list).
Cristian Iorga [Tue, 21 Aug 2012 15:21:23 +0000 (18:21 +0300)]
bitbake: Hob wizard-like reorganization
Hob is now more context-sensitive regarding
user changes/options. Also, the workflow have been
streamlined and resembles more of a wizard.
Beautified some hardcoded values.
Fixed typo.
New streamlined Image Configuration page.
Build and/or Edit image buttons presence
is context sensitive.
Recipes and packages tabs selected automatically
based on custom image or pre-defined target image
(included or all).
Context sensitive Back button.
This function takes quite a number of arguments and can be tricky to use
properly; this is not made easier if it is undocumented, so document all
of the arguments. (No functional changes, comments only.)
Richard Purdie [Tue, 21 Aug 2012 07:17:20 +0000 (08:17 +0100)]
atom-pc: Drop glibc --with-tls option, its now the only option for glibc
This option is unused by (e)glibc since 2011 and is the default. It has been
shown to interact badly with the configure option in ia32-base.inc from meta-intel
causing a rebuild of the whole system despite the only change being an
assignment with += vs =. The easiest fix is simply to drop it.
projects / thirdparty / openembedded / openembedded-core-contrib.git / log
