Recently Debian-style support for link relocation was added to
'update-alternatives' script, but it fails under circumstances where
host rootfs root directory differs from target rootfs root directory
and two alternative packages provide a symbolic link with source
located in different directories.
An example of the case is busybox provided /bin/rev (symlinking to
/bin/busybox.nosuid) and util-linux provided /usr/bin/rev (symlinking
to /usr/bin/rev.util-linux) in which case following failure occurs
during image recipe's do_rootfs() task:
ERROR: core-image-minimal-1.0-r0 do_rootfs: Postinstall scriptlets of ['util-linux'] have failed. If the intention is to defer them to first boot,
then please place them into pkg_postinst_ontarget_${PN} ().
Deferring to first boot via 'exit 1' is no longer supported.
Details of the failure are in .../tmp/work/qemux86-poky-linux/core-image-minimal/1.0-r0/temp/log.do_rootfs.
ERROR: core-image-minimal-1.0-r0 do_rootfs: Function failed: do_rootfs
Looking in log.do_rootfs file, following relevant lines can be observed:
update-alternatives: renaming rev link from /bin/rev to /usr/bin/rev
mv: cannot stat '/bin/rev': No such file or directory
Mitigate issue by applying patch which adds target root filesystem root
directory path prefix to failing 'mv' calls relevant variable references
(From OE-Core rev: f0912e23629758fe4303284e7db8f4089bb7b4cb)
Signed-off-by: Niko Mauno <niko.mauno@iki.fi>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Support explicit use of pigz when compressing the package contents by
setting:
OPKGBUILDCMD = "opkg-build -Z pigz"
(From OE-Core rev: 631e7bb7d02ca9cb9cb66e6b8395b197525cda55)
Signed-off-by: Mike Crowe <mac@mcrowe.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.
Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450). This is obviously bad.
We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.
(From OE-Core rev: ca733ba0e28d6d4c199e149ce8ae428397dfa51f)
Signed-off-by: Ross Burton <ross.burton@intel.com>
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>
0.3.5 -> 0.3.6
Patches:
- Add support for tar versions that don't support --sort
- Use local time when setting the modication time on the archives
(From OE-Core rev: fd551c5ad4c82f295470f278c524d55562a67a28)
Signed-off-by: Alejandro del Castillo <alejandro.delcastillo@ni.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We patch opkg-build to ignore hard link count changes when archiving, but in the
last upgrade (oe-core 60c9a97) the patch was refreshed to apply but the result
is that opkg-build loses the ability to switch compressor.
Refresh the patch again so that it just adds the error ignoring logic and
doesn't otherwise change the behaviour of opkg-build.
(From OE-Core rev: e8f99fa61fc00c66d216c5f3e5d37ffc42f2574b)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
To make it obvious what is lost when python is disabled, actually delete the
scripts so the user can't attempt to use them.
(From OE-Core rev: 25c8d738b904c8755df203122067afa0047b27ad)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
opkg-utils ends up in any opkg/rpm image with package management as it's the
provider for update-alternatives. If dev-pkgs is enabled then opkg-utils-dev
will get installed, which is empty but will subsequently pull python3-dev into
the image (as opkg-utils-python depends on python3).
As this can result in all of Python appearing in otherwise small images, don't
generate these pointless packages.
(From OE-Core rev: 5da7a0ba47d39612b08b4f71518b8384a3058b3f)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The scripts were fixed to be compatible with py3 some time ago,
but the shebang continued to refer to python 2.x.
(From OE-Core rev: bb5718b631151cff840bcfa171ad4f8326c2132e)
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>
0001-update-alternatives-warn-when-multiple-providers-hav.patch was
removed from SRC_URI without mention in commit 60c9a9704.
001-Makefile-use-defined-bindir-and-mandir-as-installati.patch was
removed from SRC_URI in d6b04e121.
Current opkg does not need either patch.
(From OE-Core rev: 789ec510495ca6eab379ec829e3013c977236966)
Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bump to latest revision so that update-alternatives could detect priority
conflict.
Also, we could remove the following patch because opkg-utils has already
fixed the problem in another way.
0001-Makefile-use-defined-bindir-and-mandir-as-installati.patch
[YOCTO #8314]
(From OE-Core rev: d6b04e12127dd65b96ac7f4509b829510ef21071)
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
While both result in the same in this case, postrm
should really be referring to $D as we do everywhere
else.
(From OE-Core rev: 7e459843e0371953d3d9d3ad05b019947ed7ca04)
Signed-off-by: André Draszik <adraszik@tycoint.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
--ignore-fail-on-non-empty is coreutils specific, and you
are not always going to have coreutils on target systems
especially small ones. They will use the busybox applet which
does not support --ignore-fail-on-non-empty, use pipe and true
to ignore the errorcode from rmdir instead
Fixes upgrade errors on target e.g.
rmdir: unrecognized option '--ignore-fail-on-non-empty'
BusyBox v1.24.1 (2016-12-20 10:41:39 PST) multi-call binary.
Usage: rmdir [OPTIONS] DIRECTORY...
To remove package debris, try `opkg remove update-alternatives-opkg`.
To re-attempt the install, try `opkg install update-alternatives-opkg`.
Collected errors:
* pkg_run_script: package "update-alternatives-opkg" postrm script returned status 1.
* postrm_upgrade_old_pkg: postrm script for package "update-alternatives-opkg" failed
(From OE-Core rev: 5ec8934a495e9b9a782889bc3a84936c8a381165)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If multiple providers for a utility have the same alternatives priority,
which one would be chosen is determined by which one is installed later.
Our alternatives system should be able to detect such problem and warn users
so that potential problems could be avoided.
Modify update-alternatives to warn users when detecting multiple providers
with the same priority.
[YOCTO #8314]
(From OE-Core rev: 06cf956e3441868d69f81d6c034778d855ce1c98)
(From OE-Core rev: 0b413d502868c89b7ddbe50a978317ffa774b3cc)
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When uninstalling update-alternatives, it doesn't seem to
make much sense to keep the update-alternatives database
around.
In particular when removing packaging data, e.g. due to
read-only rootfs, update-alternatives is removed from
the target file system. Leaving its database around
serves no purpose in that case as there is no way to
use it afterwards anyway.
This frees close to 700KB of (uncompressed) space in
a busybox based environment.
(From OE-Core rev: b24a63d71b517af701dfedbc7f7b541d25af708f)
Signed-off-by: André Draszik <adraszik@tycoint.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop patches now included in newer SRCREV. Update HOMEPAGE and PV
version to better align with opkg. Current revision include:
* Python 3 compatibility
* Improved error handling
(From OE-Core rev: 4b0b43f473ef12631daa577cdba39906f7d67cab)
Signed-off-by: Alejandro del Castillo <alejandro.delcastillo@ni.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This lets someone use a different update-alternatives-native provider. Without
this available, they'll step on one another in the sysroot unconditionally,
since we need to build opkg-utils-native for ipk based builds regardless.
(From OE-Core rev: 0afb74e23ee9a1fcdd334865a5f7280526785a4c)
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fix error '[[: not found' if /bin/sh is not bash.
This issue was introduced by the recent addition of tar_ignore_error.patch
to the opkg-utils recipe.
(From OE-Core rev: 2cf2242fc30b3d1ceb538ff5106e000cd888652e)
Signed-off-by: Dominic Sacré <dominic.sacre@gmx.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If a the number of hard links decreases or increases while creating the tar
files used for an ipk package, tar fails with error code 1, we use hardlinks
on package/ and packages-split/ to decrease disk usage, sometimes other
parts of the build can cause a change in the link count, since this happens
in a controlled environment we can safely ignore the error and continue to
create the ipk file, fixed some typos from old version.
[YOCTO #7933]
(From OE-Core rev: 80e6f0e5782435157b9a2687ffe6cbd3b3033b47)
Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If a the number of hard links decreases or increases while creating
the tar files used for an ipk package, tar fails with error code 1,
if this is the case we ignore the error and continue to create the ipk file
[YOCTO #7933]
(From OE-Core rev: d39e35a206c33f6062ebbaed4754a0559b28d766)
Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The new upstream commit incorporates the following fix to update-alternatives
from Chen Qi <Qi.Chen@windriver.com>:
update-alternatives: avoid double slash problem
The link path stored by update-alternatives should always be absolute and so we
do not need to add an additional slash when joining to $OPKG_OFFLINE_ROOT.
(From OE-Core rev: 49dca57255a61b05bb93dbc5d912a9d06d024058)
Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
opkg-build now checks whether tar supports the '--format' option before using
'--format=gnu' so that packages can be build with both Busybox tar (no
'--format' option) and GNU tar (defaults to posix format unless told otherwise
on some distros).
(From OE-Core rev: 99ed5ed0d2f43549e92481de388c69d65a897774)
Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Recent changes in opkg-utils allow package files to be stored in a different
directory to the package index if desired.
(From OE-Core rev: 237b9700d449de03a3f5dc524c15709f46941cf9)
Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The latest commit in opkg-utils allows packages created by opkg-build to be read
by dpkg-deb again.
(From OE-Core rev: 219944af2700ce9dbc425fac384cd32b0a802123)
Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* now with update-alternatives-cworth completely gone should correctly
replace it on target as well
(From OE-Core rev: 463e72e3125cc28a1494b42219808e925fdfe3d8)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This code makes no sense, native.bbclass clears PACKAGES anyway. Drop
it.
(From OE-Core rev: 9c8d9781794ed0886a79c8ce4544ba98be0ff858)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In small configurations its useful not to have python dependencies. This
patch adds code to disable those using PACKAGECONFIG. This allows us to
fix poky-tiny after the recent move of update-alternatives to opkg-utils.
(From OE-Core rev: 0c0039048fd97cdd3cf28f9e1c829a716dfd68ea)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Catch some u-a-cworth references that slipped through the move of u-a
to opkg-utils and its rename to -opkg.
(From OE-Core rev: a9ff0bbac5ae0688525c71f0a358f0750a277269)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The latest version of opkg-utils PROVIDES virtual/update-alternatives via a
script copied from opkg, packaged as update-alternatives-opkg.
The nativesdk class is added to the recipe to ensure that
nativesdk-update-alternatives-opkg is built.
(From OE-Core rev: 1e2c38ce13f8e4b25d8656d237343380cbc970aa)
Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
A lot of our recipes had short one-line DESCRIPTION values and no
SUMMARY value set. In this case it's much better to just set SUMMARY
since DESCRIPTION is defaulted from SUMMARY anyway and then the SUMMARY
is at least useful. I also took the opportunity to fix up a lot of the
new SUMMARY values, making them concisely explain the function of the
recipe / package where possible.
(From OE-Core rev: b8feee3cf21f70ba4ec3b822d2f596d4fc02a292)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It's not necessary to specify the protocol parameter when it's the
default protocol for the fetcher, e.g. the default protocol for
git fetcher it git, "protocol=git" isn't needed.
(From OE-Core rev: a2bab241c64428d5109c3c5ac5de4463fbad70c5)
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
On systems with dash as /bin/sh there were failures in do_package_write_ipk
due to bashisms in opkg-build. This newer revision contains a fix for this.
(From OE-Core rev: 0e185b86e80989a7904292f5866540a049cc4daa)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.
[YOCTO #3297]
(From OE-Core rev: bb67ddeb2eed3e25c626a279ef53a7e8c7bfe6f2)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* there are 2 small fixes
python-2.6 compatibility
missing C option for opkg-build
(From OE-Core rev: f7caea892cf66236bae854dcbaffec9d4a2d09a2)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
