* with usrmerge ${base_bindir} is /usr/bin/sh and libsmi fails with:
ERROR: libsmi-0.5.0-r0 do_package_qa: QA Issue: /usr/bin/smistrip contained in package libsmi requires /usr/bin/sh, but no providers found in RDEPENDS_libsmi? [file-rdeps]
ERROR: libsmi-0.5.0-r0 do_package_qa: QA run found fatal errors. Please consider fixing them.
* this is a bit strange as busybox does install /usr/bin/sh file:
./1.33.0-r0.usrmerge/image/usr/bin/sh
./1.33.0-r0.default/image/bin/sh
but also adds /bin/sh to RPROVIDES_${PN} for usrmerge in DISTRO_FEATURES:
RPROVIDES_${PN} += "${@bb.utils.contains('DISTRO_FEATURES', 'usrmerge', '/bin/sh', '', d)}"
added in:
https://git.openembedded.org/openembedded-core/commit/?id=1f6c14939c8daa5e09103789c3ff5031cc888d16
and 1.33.0-r0.usrmerge/pkgdata/runtime/busybox:FILERPROVIDES_/usr/bin/busybox.nosuid_busybox
correctly contains /usr/bin/sh
* adding runtime dependency on busybox as /usr/bin/sh provider would silence the QA warning
but people might not like adding busybox to their image
* using /bin/sh silences the QA warning as well, because /bin/sh is explicitly ignored in file-rdeps insane.bbclass:
https://git.openembedded.org/openembedded-core/tree/meta/classes/insane.bbclass?id=1f6c14939c8daa5e09103789c3ff5031cc888d16#n907
if "file-rdeps" not in skip:
ignored_file_rdeps = set(['/bin/sh', '/usr/bin/env', 'rtld(GNU_HASH)'])
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 08fad9f3a6)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
* today I've found 2 jenkins jobs stuck way too long sitting in this do_configure
Bitbake still alive (5000s)
Bitbake still alive (10000s)
Bitbake still alive (15000s)
Bitbake still alive (20000s)
Bitbake still alive (25000s)
Bitbake still alive (30000s)
... manually killed, the CMake ...
ERROR: ceres-solver-1.14.0-r0 do_configure: Execution of 'ceres-solver/1.14.0-r0/temp/run.do_configure.39438' failed with exit code 143:
...
| -- Detected Ceres being used as a git submodule, adding commit hook for Gerrit to: ceres-solver/1.14.0-r0/git/.git
| ceres-solver/1.14.0-r0/temp/run.do_configure.39438: line 213: 39485 Terminated cmake -G 'Ninja' -DCMAKE_MAKE_PROGRAM=ninja ...
I've seen it with dunfell and gatesgarth, but master has the same
ADD_GERRIT_COMMIT_HOOK function (just in newer ceres-solver release),
so probably needs the same.
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Reused below test suites from neon source package:
BASIC_TESTS:
auth
basic
request
session
socket
string-tests
stubs
uri-tests
util-tests
DAV_TESTS:
acl3744
lock
oldacl
props
xml
xmlreq
Overall execution time of above test suite is approximately 15sec.
Signed-off-by: Neetika.Singh <Neetika.Singh@kpit.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
getcwd() conforms to POSIX.1-2001 which leaves the behaviour when the
buf argument is NULL, undefined. This makes gcc 10+ throw the following
warning:
argument 1 is null but the corresponding size argument 2 value is 4096
Initially, this was fixed by disabling NSS_ENABLE_WERROR. This patch
re-enables NSS_ENABLE_WERROR (by leaving it to its default value) and
takes advantage of the existing functionality in nss that wraps the
getcwd call into a function making sure that the buf argument is always
properly allocated.
Signed-off-by: Andrei Gherzan <andrei.gherzan@huawei.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit dad2aef6be)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
The configure script contains hardcoded lookup paths to /usr and other
paths that might interfere with the host. These are overwritten with the
staging dir locations for Poky compatibility.
Signed-off-by: Anatol Belski <anbelski@linux.microsoft.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 74b66d1911)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
* I enabled the shared library option to eliminate linker issues
with a package that used spdlog.
* Edited the tweak.h file so programs use the external fmt library
headers, not the oners in spdlog.
Signed-off-by: Philip Balister <philip@opensdr.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 010cd128e7)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
This works fine until valgrind is installed on build host but its not
selected as option to build in recipe, configure wrongly pokes at
/usr/lib and sees valgrind there and enables it but only to fail in
compile time where it does not find valgrind.h in recipe sysroot, since
OE's build environment add right guardrails to pkgconfig, there is no
need to add build staging area to pkgconfig search path which turns out
to be wrong for cross builds anyway
Fixes
../../../pidgin-sipe-1.25.0/src/core/sipe-cert-crypto-nss.c:34:10: fatal error: 'valgrind.h' file not found
^~~~~~~~~~~~
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 99bc440104)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
When multilib is enabled, tmpfiles.d is created in /usr/lib rather than
${libdir} which is /usr/lib64. It causes installed-vs-shipped error:
| ERROR: colord-1.4.4-r0 do_package: QA Issue: colord:
Files/directories were installed but not shipped in any package:
| /usr/lib/tmpfiles.d
| /usr/lib/tmpfiles.d/colord.conf
Fix the path in FILES to fix the issue.
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit f147cbab5d)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
The Debian pool URLs are not permanent, and the package is no longer
included in any active Debian release. It is also not available through
any of the default Debian snapshot mirrors. Instead of adding a new
Debian snapshot mirror, use the upstream URL directly.
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 5265de84b3)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
The Debian pool URLs are not permanent and the package might vanish as
soon as the version is no longer part of any Debian release. Use the
upstream URL directly.
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 2e1a32a474)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Commits e2180b00b3 and 8edd760e66 added support for
native builds for the opensc and pcsc-lite recipes, but building
opensc-native fails after commit 40b3a51231 (2019-12-04,
"opensc: fix RDEPENDS in pcsc PACKAGECONFIG") with:
ERROR: Required build target 'opensc-native' has no buildable providers.
Missing or unbuildable dependency chain was: ['opensc-native', 'pcsc-lite-lib-native']
The commit in question is correct for target builds, but native builds
don't have packages, therefore there is no pcsc-lite-lib-native package
to depend on – the -lib part is also provided in pcsc-lite-native.
Ideally we would fix this in the opensc recipe. However, using syntax
like "PACKAGECONFIG_class-native[pcsc]" in the opensc recipe is
apparently not possible to overwrite the dependency for a native build,
and using RDEPENDS_remove has no effect either – apparently dependencies
from PACKAGECONFIG are added after RDEPENDS_remove is evaluated.
Therefore let pcsc-lite provide the missing package name for native
builds, even if fixing this unrelated package is not the most elegant
solution.
Fixes: 40b3a51231 (2019-12-04, "opensc: fix RDEPENDS in pcsc PACKAGECONFIG")
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
This reverts commit e5d9133800.
It exists in oe-core/master, but not in oe-core/gatesgarth, so this shouldn't
be included in meta-oe/gatesgarth.
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The numactl recipe now exists in openembedded-core, as of commit
dda815792b0939b2f233aa3ec7c140fc114a37d1 to oe-core. It was moved to
facilitate the requirement rt-tests now has on libnuma.
Signed-off-by: Randy Witt <randy.e.witt@linux.intel.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
One of the ptest was failing as it was missing an RDEPENDS
for its test on documentation.
Signed-off-by: Sakib Sajal <sakib.sajal@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This bugfix release contains a couple fixes for C++ and Python bindings.
Signed-off-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Latest musl shows build failures e.g.
../git/src/profile-handler.cc:275:8: error: no member named '_sigev_un' in 'sigevent'
sevp._sigev_un._tid = sys_gettid();
~~~~ ^
Signed-off-by: Khem Raj <raj.khem@gmail.com>
-License-Update(COPYING):The changes are as follows
version 0.4.9: The cidToUnicode, nameToUnicode and unicodeMap data files
installed by the poppler-data package are under the COPYING.gpl2 license
version 0.4.10: The cidToUnicode, nameToUnicode and unicodeMap data files
installed by the poppler-data package are licensed under the GNU General
Public License (GPL), version 2 (COPYING.gpl2) or version 3 (COPYING.gpl3).
-License-Update(COPYING.adobe):Copyright year updated to 2019.
Signed-off-by: Zang Ruochen <zangrc.fnst@cn.fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Emacs-minimal needs disp-table at load time, add it to the minimal package
Signed-off-by: Alejandro Enedino Hernandez Samaniego <alejandro@enedino.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Add GPLv2 into LICENSE, since this software includes alt-cannadic/COPYING which indicates GPLv2.
Signed-off-by: Taisei Nakano <taisei.nakano@miraclelinux.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This includes the following changes:
d34077d update version to v2.2.4
ffdb54c changelog: update to v2.2.4
c552ec7 mmio: clear handle ptr on error to prevent spurious munmap() in mmio_close()
c9bce3a serial: clear handle fd on error to prevent spurious close() in serial_close()
5081fe1 spi: clear handle fd on error to prevent spurious close() in spi_close()
879ef69 i2c: clear handle fd on error to prevent spurious close() in i2c_close()
711e639 gpio: clear handle fd on error to prevent spurious close() in gpio_close()
8988d25 gpio: move handle fd clear after their respective close() in gpio_close()
85a71a1 update version to v2.2.3
d8e68ab changelog: update to v2.2.3
5ecec64 spi: fix formatted bits per word truncation in spi_tostring()
5b81b89 gpio: disable cdev implementation under older gpio-cdev linux abi
3279138 make: add test for cdev gpio support in kernel headers
1de776e update version to v2.2.2
1b1df4d changelog: update to v2.2.2
50fcd0a docs/gpio: add note about kernel headers required for cdev support
05262e6 gpio: add conditional compilation of cdev implementation
caadb46 cmake: add build option for tests
ea1e0da cmake: fix identifying COMMIT_ID within build system
76b7c22 gpio: increase feature test macro version for missing definitions
80bc63d cmake: use relative source and binary directory paths
Signed-off-by: Pierre-Jean Texier <pjtexier@koncepto.io>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The multipath-tools libdmmp/Makefile install target uses 'perl -i' instead
of 'sed -i' for string substitutions. The perl method creates a temporary file
and overwrites the original which changes the inodes and corrupts the pseudo
db. Changes to pseduo cause a build abort rather than allow possible bad
ownership or permissions settings on the files. 'sed -i' is compatible
with pseudo.
see https://wiki.yoctoproject.org/wiki/Pseudo_Abort
Signed-off-by: Geoff Parker <geoffrey.parker@arthrex.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Martin Jansa
Oleksiy Obitotskyy
Aditya.Tayade
Andrei Gherzan
akuster
zhengruoqin
zangrc
Anatol Belski
Philip Balister
Khem Raj
Kai Kang
Roland Hieber
Joe Slater
Hongxu Jia
Randy Witt
Sakib Sajal
Bartosz Golaszewski
richard
changqing.li@windriver.com
Alejandro Hernandez Samaniego
Daniel Ammann
Gianfranco Costamagna
Ulrich Ölmann
Pascal Bach
Taisei Nakano
Andrej Valek
Gianluca Pacchiella
Pierre-Jean Texier
Andreas Müller
Geoff Parker