Age | Commit message (Collapse) | Author |
|
For dpkg EABI system architecture is called armel, not arm.
Reference: http://wiki.debian.org/ArmEabiPort
Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
In order to make the user experience better we now use bb.warn
on non-fatal QA errors and bb.error on fatal errors. Also note
that the RPATH problem is a security issue. (not part of the acks,
but suggested on the ML).
Signed-off-by: Tom Rini <tom_rini@mentor.com>
Acked-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Denys Dmytriyenko <denys@ti.com>
Acked-by: Eric Benard <eric@eukrea.com>
|
|
Make it clear that what's missing is an 'inherit gettext' in a
recipe.
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>
|
|
get_dependencies() searched for a filename key in dependencies
directory with the /lib/modules/kernel_version prefix while
the directory had entries stripped of it.
Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Acked-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
Acked-by: Khem Raj <raj.khem@gmail.com>
|
|
This functionality logically belongs in its own function, and it makes it
easier to explicitly run against subdirs in certain cases.
Signed-off-by: Chris Larson <chris_larson@mentor.com>
Acked-by: Khem Raj <raj.khem@gmail.com>
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
Acked-by: Khem Raj <raj.khem@gmail.com>
|
|
A function was called from ${@} in OVERRIDES which was supposed to return the
bits to prepend when in virtclass context, yet it *set* overrides instead,
returning None. This resulted in 1) adding an extra virtclass-native to
OVERRIDES each time it was expanded, and 2) appending None, causing
'localNone' to be in overrides rather than the expected 'local'.
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
* currently do_rootfs fails when it's "none", because that test returns
exit code '1':
| log_check: Using /OE/tmpdir-shr/work/spitz-oe-linux-gnueabi/initramfs-kexecboot-klibc-image-1.0-r0/temp/log.do_rootfs.21324 as logfile
| Logfile is clean
| + rm -rf /OE/tmpdir-shr/rootfs/initramfs-kexecboot-klibc-image-tmp
| + '[' none '!=' none ']'
NOTE: package initramfs-kexecboot-klibc-image-1.0-r0: task do_rootfs: Failed
ERROR: Function 'do_rootfs' failed (see /OE/tmpdir-shr/work/spitz-oe-linux-gnueabi/initramfs-kexecboot-klibc-image-1.0-r0/temp/log.do_rootfs.21324 for further information)
ERROR: Task 1356 (/OE/dev/recipes/images/initramfs-kexecboot-klibc-image.bb, do_rootfs) failed with exit code '1'
ERROR: '/OE/dev/recipes/images/initramfs-kexecboot-klibc-image.bb' failed
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
PKGR already contains DISTRO_PR, so no need to add it twice.
Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Both package_deb.bbclass and rootfs_deb.bbclass use apt-get.
Without apt system cannot be bootstrapped because there is no optimal
way to fetch new packages. Apt depends on dpkg, so switching from dpkg
to apt in EXTRA_DEPENDS provides both.
Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
Using PV/PR causes problems like libncursesw5_5.7-r16 depending on
libtinfo5 (>= 5.7+20110115).
Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
This function generates feed configuration for opkg and as such
should be used only for images using ipk, not rpm or deb.
Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
This was dangerous, as an oe_run executed from the configuration metadata
would result in caching the config metadata information, rather than the final
metadata information. We could also cache it by the datasmart instance, but I
don't think its worth the trouble at this time, particularly given the next
bitbake release exports all exported vars into the process environment
automatically, avoiding the need for any of this.
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
OE generally tries to use /var/dpkg, but the change is missing
in some places, so it is needed to move completely to one of the
options. Keeping the dpkg's default /var/lib/dpkg causes less
maintenance overhead for OE.
Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
fe706bb9af34ce43e8ec55fbf929a85d3b6d612a due to build error
* # ERROR: Error in compiling python function in:
* # ERROR: The lines leading to this error were:
* # ERROR: 22:SyntaxError:' except Exception as ex:'
|
|
Signed-off-by: Khem Raj <raj.khem@gmail.com>
|
|
Previously if ONLINE_PACKAGE_MANAGEMENT="none" but one had
FEED_DEPLOYDIR_BASE_URI set, it tried to create files in a non-existing
dir in the image (there is no /etc/opkg/ if no online package management
is requested).
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Anything that expands image to any false value should not set
INITRAMFS_TASK (instead of just "" or None).
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Mention usage of PSTAGE_MIRROR and add debugging-output for fetching.
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Tinderclient is unusable now and obsoleted by oestats-client. So we should
remove it from tree.
NOTE for Angstrom and Kaelios maintainers: this patch change your configs.
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
* fixes ERROR: Error Method already seen: git_drop_tag_prefix in
'gitver.bbclass' now in 'gitpkgv.bbclass'
Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
* A function write_dep_field() was added to do_package_ipk() to
ensure explode_deps is called for every field and to simplify
the code.
* The following statement was removed from do_package_ipk():
if not '-locale-' and not '-dbg' and not '-dev' in pkgname:
rdepends.append('%s-locale*' % pkgname)
It was added in November 2008, but can never evaluate to true.
See commit a045dfacf6f2541e0f410d9d109581f181303756
Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Acked-by: Chris Larson <chris_larson@mentor.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
* explode_deps() changed its behavior to omit version information
when the function was removed from OE in favor of BitBake's
implementation in March 2010. Since then, packages didn't contain
versioned runtime dependencies.
See commit 89b7e433719f43f1c36c76cb8856d559014e99bc
* This patch restores the previous implementation of explode_deps(),
thus fixing the generation of versioned runtime dependencies.
* Reimplementing explode_deps() using bb.utils.explode_dep_versions()
didn't work, because it choked upon parsing inline python code, e.g.
on update-modules_1.0.bb's RDEPENDS_${PN} field.
Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Acked-by: Chris Larson <chris_larson@mentor.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
* Variable 'sane' is boolean, so adding to it breaks the error condition
* Fixes previous commit 6aa31074 ported from Yocto
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Acked-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
This should fix disabling oestats with following message:
WARNING: oestats: error sending task (exception <type
'exceptions.KeyError'>, value 'qalog'), disabling stats
Signed-off-by: Yuri Bushmelev <jay4mail@gmail.com>
|
|
Do not use bb.fetch.urldata in do_distribute_sources() for traversing
SRC_URIs like was already done for do_unpack(). See 661658f62b for
additional info.
This fixes race issue with lzma-native:
http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-October/026076.html
Signed-off-by: Yuri Bushmelev <jay4mail@gmail.com>
Acked-by: Chris Larson <chris_larson@mentor.com>
Acked-by: Andrea Adami <andrea.adami@gmail.com>
|
|
oe.types.value -> oe.data.typed_value
This name has been bugging me. This function is the primary interface to
the module for OE metadata, as it takes a variable name and datastore and
returns an object of the correct type. While this function is part of the
variable typing implementation, in reality it's more about giving you a
useful object from the metadata, so I think oe.data is a more appropriate
place for it.
oe.types -> oe.maketype
These are the functions which construct types, not the types themselves,
so it was somewhat misleading.
oe._types -> oe.types
These are the actual types, any callable in this module becomes an OE
type, using its arguments to determine the variable flags (optional and
required) to obey. Will use __init__'s args in the case of an actual
python class.
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
We need the javah from icedtea6 here.
|
|
After fixing update-alternatives-native the command:
rmdir ${IMAGE_ROOTFS}/var/dpkg/alternatives
fails, because the directory is not empty, and the command:
ln -s ${libdir}/opkg/alternatives ${IMAGE_ROOTFS}/var/dpkg/alternatives
creates a dangling symlink, because we mkdir only the parent directory of
the target.
Fixes: mkdir -p the alternatives directory itself and move alternatives
definitions into it before rmdiring the original alternatives directory
Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Bitbake's pyshlex.py cannot parse it raising NotImplementedError in
line 295 (bitbake @ c8d8388ce9662574a390be3b849c07d1b044275d).
Signed-off-by: Filip Zyzniewski <filip.zyzniewski@gmail.com>
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
|
|
* Relax empty data checking (replace asserts)
* Replace direct path manipulations with os.path.join
* Replace bb.note() with bb.warn() where appropriate
* Improve diagnostic messages
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
* This patch adds an empty compileconfigs function to the
sourceipk class.
* This is required because when making a sourceipk of a
package using the multi-kernel.inc functionality you will
have a race condition between the compileconfigs of the
multi-kernel file building the kernel and sourceipk trying
to copy and patch the sources. The resulting conditions
can then occur:
- You will package some interim build files, yielding
a sourceipk that is not consistent between builds
- You will try to copy interim build files as they are
being deleted which will result in an error and break
the build.
* The empty compileconfigs function is overwritten by the
real compileconfigs function when the multi-kernel.inc file
is used.
Signed-off-by: Chase Maupin <Chase.Maupin@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
|
|
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
* Use it as a hint for BitBake to know that packages inheriting
kernel.bbclass do provide kernel-x.y.z-localversion.
Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Acked-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: Paul Menzel <paulepanter@users.sourceforge.net>
for `omap3-sgx-modules` with `angstrom-2008.1`
|
|
* bump EFL_SRCREV a bit to current HEAD
* introduce EFL_SRCREV_1.0.0 for recipes which don't need SRCREV change
so often like stuff from OLD and BROKEN dirs and themes
* move SRC_URI and S settings from bbclass to svn recipes
* replace E_CVS and E_URI with E_SVN and E_MIRROR
* PV/PR will be updated in following patches for individual recipes
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
bitbake 1.8.x has a broken bb.plain. This means that our build
simply blows up before we can tell the user to upgrade to BB_MIN_VERSION.
Reported by Christian Poessinger.
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
${KERNEL_VERSION})
* As a special case, kernel modules must depend on
kernel-${KERNEL_VERSION}, which gets runtime provided by the kernel
package. The reason is, that the package version can be different
from ${KERNEL_VERSION}, because the latter depends on the kernel
configuration, e.g. CONFIG_LOCALVERSION.
Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
|
|
Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
|
|
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
Aside: should really shift most of the bits from this class into a python
module in the oe package, which would make it possible to manipulate the data
from the metadata if necessary.
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|