Patrick Williams [Thu, 15 Sep 2016 21:11:16 +0000 (16:11 -0500)]
util-linux: Disable minix support.
Disable building of {mkfs,fsck}.minix, which are used
to support minix file-systems. Minix predates Linux and
support for its file-systems is unlikely to be needed.
No recipes otherwise reference minix, except in patches
to autotools configuration scripts, so there should be no
impact to other recipes.
By default cmake will auto-detect if a library is present on the host and if it
isn't present will use an internal fork. For some libraries using the internal
fork is preferable as it can be built with less dependencies, but for others
we're either already building it or the impact of building it is comparable to
internal build.
Continue to use the internal fork of libarchive as our libarchive-native has a
large number of build dependencies. Using the internal libarchive means that
system bzip2 and zlib must be used.
Explicitly use the internal fork of jsoncpp as we don't have this in oe-core.
Explicitly depend on curl-native, expat-native, and xz-native to ensure these
dependencies are not floating. curl-native is a non-trivial dependency but is
comparable to building the internal fork, so there's no reason to build it
twice.
Change bzip2-native to bzip2-replacement-native as bzip2-native is
ASSUME_PROVIDED.
Markus Lehtonen [Wed, 14 Sep 2016 11:24:42 +0000 (14:24 +0300)]
scripts: introduce buildstats-diff
New script for comparing buildstats from two separate builds. The script
has two modes: normally it prints the differences in task execution
(cpu) times but using --ver-diff option makes it just print the recipe
version differences without any cpu time data. Other command line
options are provided to alter the sort criteria of the data and to
filter out insignificant differences and/or short tasks.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Previously the following commit in oe-core move RPM metadata
from DEPLOY_DIR to WORKDIR.
-----------
commit a92c196449c516fe51786d429078bbb1213bb029
Author: Stephano Cetola <stephano.cetola@linux.intel.com>
Date: Wed Aug 10 13:03:16 2016 -0700
Allow for simultaneous do_rootfs tasks with rpm
Give each rootfs its own RPM channel to use. This puts the RPM metadata
in a private subdirectory of $WORKDIR, rather than living in DEPLOY_DIR
where other tasks may race with it.
-----------
In the modification of 'class RpmIndexer, it should not
directly set arch_dir with WORKDIR. It caused 'bitbake
package-index' could not work correctly.
Assign WORKDIR as input parameter at RpmIndexer initial time
could fix the issue.
Shan Hai [Tue, 13 Sep 2016 05:45:46 +0000 (13:45 +0800)]
shadow: use relaxed usernames
The groupadd from shadow does not allow upper case group names, the
same is true for the upstream shadow. But distributions like
Debian/Ubuntu/CentOS has their own way to cope with this problem,
this patch is picked up from CentOS release 7.0 to relax the usernames
restrictions to allow the upper case group names, and the relaxation is
POSIX compliant because POSIX indicate that usernames are composed of
characters from the portable filename character set [A-Za-z0-9._-].
Paul Eggleton [Tue, 13 Sep 2016 02:15:14 +0000 (14:15 +1200)]
oeqa/sdkext/devtool: use a smaller module to test node.js functionality
The "forever" package, despite its innocent description, actually drags
in a surprising number of dependencies and as a result the nodejs test
takes up to 10 minutes as a result. Pick a different example with a much
more reasonable set of dependencies.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andre McCurdy [Mon, 12 Sep 2016 20:14:05 +0000 (13:14 -0700)]
makedevs: don't restrict device node paths to 40 characters
40 character paths work OK for device nodes in /dev but not for
device nodes created in a chroot, LXC container, etc.
Since the 'path' array is already a 4k buffer, the sscanf 40
character limit seems to be a typo or historical mis-merge. Update
the sscanf limit and bring the code in sync with the Buildroot
version:
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This logs the launch command that was used for runqemu while running -c
testimage. This way, if I'd like to easily launch qemu manually in
order to debug a failed test, I know what commmand was run to create
the qemu instance.
Ross Burton [Wed, 14 Sep 2016 16:41:35 +0000 (17:41 +0100)]
cmake: don't inherit autotools
cmake doesn't use autotools, the functions get replaced by either cmake.bbclass
(target) or the recipe itself (native) leaving just lots of superfluous
dependencies.
MIPS64 target was being configured for linux-mips which defaults to
MIPS32. Doesn't cause any issue as far as I can see but it would be
wiser to use the correct target configuration.
Ross Burton [Wed, 14 Sep 2016 23:08:01 +0000 (00:08 +0100)]
sanity: make license QA failures fatal again
Previous work to clean up the license QA code (oe-core fbdf977) had the side
effect that failing the license sanity check (bad or missing LIC_FILES_CHKSUM)
would emit an error message but wouldn't actually abort the build.
Solve this by changing populate_lic_qa_checksum() so that it tracks if the
message class was in ERROR_QA and if so, aborts the function.
Richard Purdie [Thu, 15 Sep 2016 10:48:39 +0000 (11:48 +0100)]
openjade/sgml-common: Add sstate postrm commands
If you bump the PR of sgml-common and openjade-native but not sgml-
common-native, you will see a failure as files were removed from the
sysroot but still referenced by the sgml docbook catalog.
To properly handle this, the clean function needs to run at sstate
removal time, the problem is that this sstate removal can happen when
the metadata isn't present, so the correct removal commands are
unknown.
To avoid this, we need to write the commands into a "postrm" script
when we install the files, this can then be executed at sstate removal
time.
Richard Purdie [Thu, 15 Sep 2016 10:45:14 +0000 (11:45 +0100)]
sstate: Add postrm command support
There are some issues in sstate which can't be handled by file removal
alone. Currently there is no way to execute a command against sstate and
doing so is potentially problematic for things like dependencies. This
patch adds a mechanism where any "postrm" script is executed if its present
allowing some openjade/sgml issues to be resolved.
Richard Purdie [Wed, 14 Sep 2016 21:29:43 +0000 (22:29 +0100)]
bitbake: build: Allow dirs/cleandirs to work for empty functions
Users are surprised when dirs/cleandirs aren't acted upon for
empty functions. This reorders the code slightly so that those
flags are acted upon for empty functions as there are cases where
this is expected.
bitbake: bitbake: fetch2: Make SRCREV_FORMAT name substitution safer
The implementation of SRCREV_FORMAT has at least two issues:
1. Given two names "foo" and "foobar" and SRCREV_FORMAT = "foo_foobar",
"foo" might currently get substituted twice, and "foobar" not at
all.
2. If the revision substitued for some name happens to contain another
name as a substring, then that substring might incorrectly get
replaced.
Fix both issues by sorting the names with the longest ones first and
replacing all names at once with a regular expression. This was inspired
by
http://stackoverflow.com/questions/6116978/python-replace-multiple-strings.
Paul Eggleton [Tue, 13 Sep 2016 02:11:41 +0000 (14:11 +1200)]
bitbake: fetch2/npm: don't download same URL multiple times
If we've already fetched a particular URL then we do not need to do so
again within in the same operation. Maintain an internal list of fetched
URLs to avoid doing that.
Ed Bartosh [Thu, 8 Sep 2016 11:02:33 +0000 (14:02 +0300)]
bitbake: event.py: output errors and warnings to stderr
All logging messages are printed on stdout when processing
UI event queue. This makes it impossible to distinguish between
errors and normal bitbake output. Output to stderror or stdout
depending on log level should fix this.
oeqa/utils/decorators: LogResults fix race condition in linkfile
In order to avoid race condition when test if exists the linkfile
use bb.utils.lock, the best solution is to create a unique name
for the link file.
There is no way to create a unique linkfile name at this decorator
because is needed the machine and image variables, those variables
can't be passed easily in this code.
To avoid broke test export functionality use a try/except because bb
isn't available when use test export
Andre McCurdy [Mon, 12 Sep 2016 22:57:45 +0000 (15:57 -0700)]
image.bbclass: add IMAGE_DEVICE_TABLE to rootfs[vardeps]
IMAGE_DEVICE_TABLE and IMAGE_DEVICE_TABLES are both referenced by
_create_devfs, therefore ensure that rootfs is rebuilt if changes
are made to either variable.
Ed Bartosh [Wed, 14 Sep 2016 12:02:04 +0000 (15:02 +0300)]
create-pull-request: remove output directory
When 'git request-pull' fails it makes sense to remove output
directory. Otherwise create-pull-request will complain that
output directory already exists on the next run.
* Add patch Skip-empty-section-fixes-66.patch to prevent errors like:
/
|ERROR: go-cross-1.6.2-r0 do_populate_sysroot_setscene: '('patchelf-uninative',
|'--set-interpreter', '/home/user/src/prj/build/tmp/sysroots-uninative/
|x86_64-linux/lib/ld-linux-x86-64.so.2', '/home/user/src/prj/build/tmp/
|work/x86_64-linux/go-cross/1.6.2-r0/sstate-install-populate_sysroot/x86_64-
|linux/usr/bin/aarch64-prj-linux/go')'
|failed with exit code 1 and the following output:
|cannot find section .rela.dyn
\
* Add patch handle-read-only-files.patch to fix error when building
eSKD, the following error appears on task do_testsdkext
/
|ERROR: db-native-6.0.30-r1 do_populate_sysroot_setscene: '('patchelf-uninative',
|'--set-interpreter', 'src/fsl-community-bsp/build/tmp/work/qemuarm-poky-linux-
|gnueabi/core-image-minimal/1.0-r0/testsdkext/tc/tmp/sysroots-uninative/
|x86_64-linux/lib/ld-linux-x86-64.so.2', 'src/fsl-community-bsp/build/tmp/work/
|qemuarm-poky-linux-gnueabi/core-image-minimal/1.0-r0/testsdkext/tc/tmp/work/
|x86_64-linux/db-native/6.0.30-r1/sstate-install-populate_sysroot/x86_64-linux/
|usr/bin/db_tuner')' failed with exit code 1 and the following output:
|b'open: Permission denied\n
\
* Add patch Increase-maxSize-to-64MB.patch to fix error described
bellow, the same issue is discussed here:
- https://github.com/NixOS/patchelf/issues/47
/
|ERROR: qemu-native-2.5.0-r1 do_populate_sysroot_setscene: '('patchelf-uninative',
|'--set-interpreter', '../build/tmp/sysroots-uninative/x86_64-linux/lib/
|ld-linux-x86-64.so.2', '../build/tmp/work/x86_64-linux/qemu-native/2.5.0-r1/
|sstate-install-populate_sysroot/x86_64-linux/usr/bin/qemu-mips64')'
|failed with exit code 1 and the following output:
|warning: working around a Linux kernel bug by creating a hole of 36032512
|bytes in ‘../build/tmp/work/x86_64-linux/qemu-native/2.5.0-r1/
|sstate-install-populate_sysroot/x86_64-linux/usr/bin/qemu-mips64’
|maximum file size exceeded
\
When building busybox, an occasional error was observed.
The error is consistently the same:
libbb/appletlib.c:164:13: error: 'NUM_APPLETS' undeclared (first use in this function)
while (i < NUM_APPLETS) {
The reason is the include file where NUM_APPLETS is defined is not yet generated (or is being modified)
at the time libbb/appletlib.c is compiled.
The attached patchset fixes the problem by assuring libb is compiled as the last directory.
Robert Yang [Wed, 14 Sep 2016 09:45:57 +0000 (02:45 -0700)]
runqemu: fix a race issue on lockdir
There might be a race issue when multi runqemu processess are
running at the same time:
| Traceback (most recent call last):
| File "/home/pokybuild/yocto-autobuilder/yocto-worker/nightly-ipk/build/scripts/runqemu", line 920, in <module>
| ret = main()
| File "/home/pokybuild/yocto-autobuilder/yocto-worker/nightly-ipk/build/scripts/runqemu", line 911, in main
| config.setup_network()
| File "/home/pokybuild/yocto-autobuilder/yocto-worker/nightly-ipk/build/scripts/runqemu", line 760, in setup_network
| self.setup_tap()
| File "/home/pokybuild/yocto-autobuilder/yocto-worker/nightly-ipk/build/scripts/runqemu", line 697, in setup_tap
| os.mkdir(lockdir)
| FileExistsError: [Errno 17] File exists: '/tmp/qemu-tap-locks'
Markus Lehtonen [Wed, 14 Sep 2016 08:46:06 +0000 (11:46 +0300)]
python-smartpm: use md5 as the digest for rpm_sys channel
Use md5 sum instead of mtime as the "digest" method for rpm_sys channel.
The digest is used to determine if the channel has been updated. It was
found out that mtime was not a reliable digest. On some systems mtime
of the rpm db does not get updated after every transaction if transactions
(smart install / remove commands) are fired in quick succession. As a
consequence smartpm cache and rpm db get out of sync.
Paul Eggleton [Wed, 14 Sep 2016 05:09:52 +0000 (17:09 +1200)]
lib/oe/copy_buildsystem: fix building eSDK with indirect paths in BBLAYERS
Indirect paths (e.g. ${TOPDIR}/../meta-something) do generally work if
used in BBLAYERS in bblayers.conf. However, if you built an extensible
SDK with this configuration then the creation of the workspace within
the SDK using devtool in do_populate_sdk_ext failed. This is because
the copy_buildsystem code was no longer correctly recognising that the
core layer ("meta") was part of a repository (e.g. openembedded-core /
poky) that should be shipped together - because of the indirection - and
thus it was splitting out the meta directory, and a number of places in
the code assume that the meta directory is next to the scripts
directory. Use os.path.abspath() to flatten out any indirections.
Paul Eggleton [Wed, 14 Sep 2016 05:09:51 +0000 (17:09 +1200)]
classes/populate_sdk_ext: ensure we clean the right temporary TMPDIR path
After we run the build system within the eSDK internally as part of the
sstate filtering that happens during do_populate_sdk_ext, we need to
ensure that the TMPDIR created during that process gets deleted. However
we were using the TMPDIR path for the build producing the eSDK which may
not be the same (since that value would typically be filtered out) thus
if the user had set TMPDIR to something other than the default, the
temporary TMPDIR would not be deleted which not only led to extraneous
junk entering the SDK but also failures during install because the
TMPDIR path was different. In order to fix this, force TMPDIR to a known
value during the sstate filtering run so we know what to delete
afterwards.
Richard Purdie [Tue, 13 Sep 2016 13:56:48 +0000 (14:56 +0100)]
eudev: Add missing pkgconfig dependencies
Building eudev from scratch in an sstate build results in configure errors
since the pkg-config macros can't be found. Add in a missing pkg-config
dependency using the appropriate class to avoid such failures.
Richard Purdie [Tue, 13 Sep 2016 13:56:15 +0000 (14:56 +0100)]
autotools: Correctly filter m4 files view to setscene dependencies
Currently when you run builds from sstate, you can see warnings like:
WARNING: systemd-1_230+gitAUTOINC+3a74d4fc90-r0 do_configure: /data/poky-master/tmp-glibc/sstate-control/manifest-intel-corei7-64-glibc-initial.populate_sysroot not found
WARNING: systemd-1_230+gitAUTOINC+3a74d4fc90-r0 do_configure: /data/poky-master/tmp-glibc/sstate-control/manifest-intel-corei7-64-libgcc-initial.populate_sysroot not found
This is due to co_configure wanting to copy a limited number of m4 macros,
only listed in a recipes DEPENDS but that set is still larger than the set of
recipes which get restored from sstate.
For build determinism and to avoid these warnings, we need to make this
function match what the sstate code does. We really don't want to duplicate
the functionality since keeping things in sync would be hard so we create
a data structure which can be passed into the same underlying function,
setscene_depvalid().
André Draszik [Tue, 13 Sep 2016 09:07:10 +0000 (10:07 +0100)]
kernel-yocto: remove do_shared_workdir from SRCTREECOVEREDTASKS
Various recipes depend on the kernel's do_shared_workdir
task, a quick grep suggests all external kernel modules
(via module.bbclass), but also perf, and potentially any
additional headers as outlined in linux-libc-headers.inc
are affected.
Having do_shared_workdir in SRCTREECOVEREDTASKS means this
task is removed when externalsrc is enabled, making all
those recipes fail as the task they depend on,
virtual/kernel:do_shared_workdir, doesn't exist.
Remove do_shared_workdir from SRCTREECOVEREDTASKS so that
all those recipes work even if externalsrc is activated.
According to the comment in here, the reason for
do_shared_workdir to be removed as a task is because it
modifies the source tree, but that doesn't seem to be
case.
André Draszik [Mon, 5 Sep 2016 08:33:54 +0000 (09:33 +0100)]
image.bbclass: also mask debugfs versions of masked images
When IMAGE_GEN_DEBUGFS is enabled, and IMAGE_FSTYPES_DEBUGFS is left
at its default (as suggested by local.conf.sample.extended),
recipe parsing fails:
bitbake kern-tools-native # or anything else for that matter
ERROR: <poky.git>/meta/recipes-core/images/build-appliance-image_15.0.0.bb: No IMAGE_CMD defined for IMAGE_FSTYPES entry 'debugfs_vmdk' - possibly invalid type name or missing support class
ERROR: Failed to parse recipe: <poky.git>/meta/recipes-core/images/build-appliance-image_15.0.0.bb
Summary: There was 1 WARNING message shown.
Summary: There were 2 ERROR messages shown, returning a non-zero exit code.
i.e. bitbake doesn't even finish parsing...
Since IMAGE_FSTYPES_DEBUGFS is based on IMAGE_FSTYPES, and
since the build-appliance-image is setting IMAGE_FSTYPES
to vmdk, image.bbclass/image_types.bbclass will be trying
to build a debugfs_vmdk, causing the error, as this is not
implemented.
One solution to solving this problem could be as simple as
adding a line
IMAGE_FSTYPES_DEBUGFS_remove = "vmdk"
to the build-appliance-image recipe, but that is very
specific to the error encountered and carries the risk of
the error being reintroduced in another recipe.
Another solution could be to add 'debugfs_vmdk' to
IMAGE_TYPES_MASKED in image-vm.bbclass, but again, this
approach doesn't seem generic enough.
None of the live and vm type images have an implementation
for building a debugfs version, it doesn't seem to make
sense to build debugfs versions of any of them anyway, and
given IMAGE_TYPES_MASKED appears to be intended for those
image types exclusively, it seems the right approach is to
unconditionally also mask all debugfs_ flavours from
IMAGE_TYPES_MASKED to achieve a generic solution.
Signed-off-by: André Draszik <git@andred.net> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
André Draszik [Mon, 5 Sep 2016 08:34:42 +0000 (09:34 +0100)]
image.bbclass: do not check size of the debugfs image
The debugfs is supposed to be used in addition to the
normal image for debugging purposes, it doesn't make
sense to artificially limit its maximum size.
Signed-off-by: André Draszik <git@andred.net> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Max Krummenacher [Sun, 11 Sep 2016 17:13:20 +0000 (19:13 +0200)]
sstate.bbclass: normalize whitelisted paths
Without this a whitelisted path containing /../ breaks the test for a file
allowed to be provided by more than one recipe.
Noticed when local.conf contains:
DEPLOY_DIR = "${TOPDIR}/../deploy"
|ERROR: core-image-minimal-1.0-r0 do_image_complete: The recipe
| core-image-minimal is trying to install files into a shared area when those
| files already exist. Those files and their manifest location are:
| .../poky/deploy/images/qemux86/README_-_DO_NOT_DELETE_FILES_IN_THIS_DIRECTORY.txt
| Matched in b'manifest-qemux86-linux-yocto.deploy'
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
classes/populate_sdk_ext: exclude image tasks from locked signatures
Tasks for image recipes cannot be locked and should be excluded from eSDK
generated locked-sigs.inc. get_sdk_install_targets() was not returning right
image targets to be excluded incase of 'minimal' sdk. This change fixes the issue.
By letting a recipe assign GI_DATA_ENABLED trivially there is a simple and clear
way to disable gobject-introspection for specific build configurations.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
gstreamer1.0-plugins-bad: Fix packageconfigs and patches
This reintroduces fixes and changes which were introduced in the original
gstreamer1.0-plugins-bad 1.8.1 upgrade commit.
* packageconfigs changed since GStreamer 1.6.3
(they often do between minor version increases like 1.6 -> 1.8)
* hls,tinyalsa packageconfigs moved into the .inc file
* vulkan packageconfig dropped since there are no vulkan libraries in OE
(libxcb alone is not enough)
* reintroduced glimagesink downrank patch (it was removed because it was
dangling before)
* fixed patch line numbers
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Maxin B. John <maxin.john@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When trying to re-test smart rpm tests. A fail arises
due to a linux user previously created upon the image.
We've added a few lines to delete such user and his home dir
when finishing test.
Markus Lehtonen [Thu, 8 Sep 2016 16:47:46 +0000 (19:47 +0300)]
oeqa.buildperf: another fix for splitting 'nevr' string
When processing buildstats we determine recipe name, epoch, version and
revision from the per-recipe buildstat directory name. One previous
patch made an assumption that package version starts with a number.
That might not be true because of a packaging mistake or whatever
reason. Thus, if a version starting with a number is not found, fall
back to the "old" method of just taking the second-last dash-delimited
part (the one before revision).
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Maxin B. John <maxin.john@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Jussi Kukkonen [Thu, 8 Sep 2016 13:35:31 +0000 (16:35 +0300)]
tiff: Update download URL
remotesensing.org domain has been taken over by someone unrelated.
There does not seem to be an up-to-date tiff homepage, but
osgeo.org is a reliable download site.
Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
libpcre: explicitely specify CCLD_FOR_BUILD for this recipe
libpcre uses CCLD_FOR_BUILD as the name of the host command to use for
linking. This is not a standard autotools symbol but particular to this
recipe. We need to set it explicitely
version 5.4.1 of rpm was not properly distclean before release, which
causes problems when cross-compiling.
The previous version this recipe called make distclean, but that would
trigger a call to ./configure which would fail when no gcc is
available and make the whole do_configure fail further down the line
This patch manually removes the files from the recipe.
busybox: allow overriding compiler name in KConfig
The KConfig infrastructure needs to build HOST binaries in order to
provide its infratstructure. Yocto needs to force the HOSTCC and HOSTCPP
variables to BUILD_CC and BUILD_CPP to make sure that the proper compiler
is used when compiling host binaries
linux-libc-headers: allow overriding compiler name in KConfig
The KConfig infrastructure needs to build HOST binaries in order to
provide its infratstructure. Yocto needs to force the HOSTCC and HOSTCPP
variables to BUILD_CC and BUILD_CPP to make sure that the proper compiler
is used when compiling host binaries
linux-yocto: allow overriding compiler name in KConfig
The KConfig infrastructure needs to build HOST binaries in order to
provide its infratstructure. Yocto needs to force the HOSTCC and HOSTCPP
variables to BUILD_CC and BUILD_CPP to make sure that the proper compiler
is used when compiling host binaries
cross: export CC family from BUILD_* for cross compilers
cross-compilers are native recipes that need to be compiled by the host's
compiler. However they do not use native.bbclass
As a consequence, the various CC, CXX etc environment variables are not
correctly set and they will not honor the host compiler name provided
by the BUILD_* variables.
Nicolas Dechesne [Fri, 26 Aug 2016 12:05:00 +0000 (13:05 +0100)]
gstreamer1.0-plugins-good: enable v4l2-probe
A new mechanism to probe v4l2 M2M devices was implemented in gst 1.8 series, in
order to get such devices probed we now need to enable v4l2-probe compile option
which upstream decided to keep disabled by default (unfortunately), see [1].
With this feature disabled, it is not possible to get v4l2 M2M device to work in
Gstreamer which is a common use case on many embedded platforms. This patch
enables this new option as soon as v4l2 support is enabled in gstreamer -good.
Signed-off-by: Nicolas Dechesne <nicolas.dechesne@linaro.org> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Marek Vasut <marex@denx.de> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joe Slater [Tue, 16 Aug 2016 18:16:20 +0000 (11:16 -0700)]
useradd: do not delete users and groups during configure
If two recipes both create the same users and groups, the
second recipe can delete items created by the first causing
things like "chown" to fail for the first recipe.
Signed-off-by: Joe Slater <jslater@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
and A has errors in parselogs, machine B can pick these up and cause
immense confusion. This is because the test transfers the log files
to cwd which is usually TOPDIR. This is clearly bad and this patch
uses a subdir of WORKDIR to ensure machines don't contaminate each
other.
Also ensure any previous logs are cleaned up from any existing
transfer directory.
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Markus Lehtonen [Fri, 9 Sep 2016 08:27:11 +0000 (11:27 +0300)]
oeqa.runtime.smart: work around smart race issues
Yucku hack around test failures which ultimately are caused by a race in
smartpm itself. Issuing smartpm commands in quick succession causes
races in package cache of smartpm on some systems. This patch mitigates
the problem by sleeping for 1 second after each smartpm command that
modifies the system.
Richard Purdie [Thu, 8 Sep 2016 10:24:00 +0000 (11:24 +0100)]
oeqa/runtime/smart: Prune feeds to save memory
Full package feed indexes overload a 256MB image so reduce the number of rpms
the feed. Filter to p* since we use the psplash packages and this leaves some
allarch and machine arch packages too.
Ross Burton [Mon, 5 Sep 2016 21:25:44 +0000 (22:25 +0100)]
gtk-doc: use pkg-config-native in native gtk-doc.m4
When building gtk-doc-native the m4 functions for autoconf should use
pkg-config-native instead of pkg-config so that they can find the native
tooling.
This means that it is possible to generate gtk-doc without building the target
packages.
Robert Yang [Mon, 5 Sep 2016 06:56:24 +0000 (23:56 -0700)]
gobject-introspection: set GI_SCANNER_DISABLE_CACHE for native
The native recipe should not write files to $HOME/.cache as target, this can
avoid problems when multi builds are running on the same host like:
| File "./g-ir-scanner", line 66, in <module>
| sys.exit(scanner_main(sys.argv))
| File "../gobject-introspection-1.48.0/giscanner/scannermain.py", line 543, in scanner_main
| transformer = create_transformer(namespace, options)
| File "../gobject-introspection-1.48.0/giscanner/scannermain.py", line 389, in create_transformer
| symbol_filter_cmd=options.symbol_filter_cmd)
| File "../gobject-introspection-1.48.0/giscanner/transformer.py", line 54, in __init__
| self._cachestore = CacheStore()
| File "../gobject-introspection-1.48.0/giscanner/cachestore.py", line 61, in __init__
| self._check_cache_version()
| File "../gobject-introspection-1.48.0/giscanner/cachestore.py", line 89, in _check_cache_version
| self._clean()
| File "../gobject-introspection-1.48.0/giscanner/cachestore.py", line 141, in _clean
| self._remove_filename(os.path.join(self._directory, filename))
| File "../gobject-introspection-1.48.0/giscanner/cachestore.py", line 123, in _remove_filename
| os.unlink(filename)
| FileNotFoundError: [Errno 2] No such file or directory: '/home/pokybuild/.cache/g-ir-scanner/0a47aa95823c95a0b5d1bd610b60d02f35785f26'
| Makefile:3518: recipe for target 'GModule-2.0.gir' failed
projects / thirdparty / openembedded / openembedded-core-contrib.git / log
