If BASELIB is set to non-default value (like lib64),
next 'installed but not shipped' issue happens:
/usr/lib
/usr/lib/cmake
/usr/lib/pkgconfig
/usr/lib/cmake/RapidJSON
/usr/lib/cmake/RapidJSON/RapidJSONConfigVersion.cmake
/usr/lib/cmake/RapidJSON/RapidJSONConfig.cmake
/usr/lib/pkgconfig/RapidJSON.pc
This is because rapidjson has library installation dir set
to 'lib' by default and hardcoded in another place.
Fix this by next changes:
- set appropriate LIB_INSTALL_DIR
- fix hardcoded 'lib' in CMAKECONFIG_INSTALL_DIR case
Signed-off-by: Ruslan Bilovol <rbilovol@cisco.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
In case some recipes might need it, for instance, the aktualizr-native
of meta-updater.
Signed-off-by: Ming Liu <liu.ming50@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Drop backported patch as it's present in 1.6. Switch to autotools-brokensep
to avoid
| sed -e 's/\\/\\\\/g' -e 's/"/\\"/g' -e 's/^/"/' -e 's/$/\\n"/' ../jq-1.6/src/builtin.jq > src/builtin.inc
| /bin/sh: src/builtin.inc: No such file or directory
License-Update: whitespace changes
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
A couple have still been missed in the past despite multiple
attempts at doing so (or simply have re-appeared?).
Search & replace made using the following command:
sed -e 's|\(d\.getVar \?\)( \?\([^,()]*\), \?True)|\1(\2)|g' \
-i $(git grep -E 'getVar ?\( ?([^,()]*), ?True\)' \
| cut -d':' -f1 \
| sort -u)
Signed-off-by: André Draszik <andre.draszik@jci.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
guider [1] is a system-wide performance analyzer for platforms based
on linux kernel. It can measure a variety of system resource usage and
trace thread/process activities. It also reports vary meaningful
output to analyze system status.
[1] https://github.com/iipeace/guider
Signed-off-by: Peace Lee <iipeace5@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
- Drop upstreamed ucontext_t patch
- basename is now available in musl
- Pin ARM-only ISA on armv5 and older arches
- Use COMPATIBLE_HOST instead of COMPATIBLE_MACHINE
- memory.h has been renamed to memory_allocator.h in latest breakpad
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This is in fact an update from 1.7.1 to 1.10.0. (The git hash
of the 1.9.0 recipe referred flatbuffers release 1.7.1.)
This patch is obsolete, a similar fix has been applied upstream:
- 0001-correct-version-for-so-lib.patch
These two patches have been applied upstream:
- 0001-flatbuffers-Move-EndianSwap-template-to-flatbuffers-.patch
- 0002-use-__builtin_bswap16-when-building-with-clang.patch
Signed-off-by: Adrian Freihofer <adrian.freihofer@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Drop '-f, --force' operator from rm command as that will hide the fact
if the removed artifacts cease to exist in later version of this recipe.
Also drop a superfluous slash from artifact path.
Signed-off-by: Niko Mauno <niko.mauno@iki.fi>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Add missing space characters around the operator in SRCNAME declaration
Signed-off-by: Niko Mauno <niko.mauno@iki.fi>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Remove redundant PYTHON* declarations/references as this recipe builds
now correctly without them.
Signed-off-by: Niko Mauno <niko.mauno@iki.fi>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/meta?id=1dccb37d89f54d7284b6a81957ab228906f24e1c
has replaced PYTHON_PN references with hardcoded 'python' and 'python3'
strings in pythonnative.bbclass and python3native.bbclass, respectively.
Associated commit message states: "If both are inherited then both
native Pythons are available, but the last to be inherited will be the
one to set PYTHON."
Since after the above-mentioned commit this Pythonv2 recipe installs
wrongly into /usr/lib/python3.5/ directory, move inherit of
distutils-base.bbclass (which inherits pythonnative.bbclass) to right
hand side of gobject-introspection.bbclass (which inherits
python3native.bbclass) which changes the install path back to
/usr/lib/python2.7/ mitigating the issue.
Signed-off-by: Niko Mauno <niko.mauno@iki.fi>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Doxygen binary is necessary to generate C++ bindings for libsigrok.
Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
As there is below logic in configure.in
[snip]
test -z "$TK_LIBRARY" && TK_LIBRARY='$(prefix)/lib/tk$(VERSION)'
[snip]
It can make TK_LIBRARY to be $(prefix)/lib/tk$(VERSION)
and then result in some files installed to
/usr/lib/tk8.6 even when ${libdir} actually
extract as /usr/lib64 and there is also a commit
as below adding hack to workaround this.
1ca29d1cd tk: make multilib build compatible and fix library install
Export TK_LIBRARY='${libdir}/tk${VER}' to guarantee
the files installed to the expected places and
also remove the workaround in the previous commit
1ca29d1cd.
Signed-off-by: Mingli Yu <Mingli.Yu@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
After enable tk via PACKAGECONFIG for python3,
there comes below error:
Missing or unbuildable dependency chain was: ['buildtools-tarball', 'nativesdk-python3', 'nativesdk-tk']
Extend tk to nativesdk to fix this issue.
Signed-off-by: Mingli Yu <Mingli.Yu@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Causes the empty base package to be created so that RapidJSON can be
sanely included in an SDK via IMAGE_INSTALL, RDEPENDS, and the like.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
If TARGET_ARCH is not set to i386 for x86, build will be failed
with logs such as "arch/i586: No such file or directory".
In this case, TARGET_ARCH is set to i586.
To succeed the build, add additional function named set_target_arch.
Signed-off-by: Yi-Soo An <yisooan@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Bring in a newer version of rapidjson which includes a fixes when
running on ARM, specifically 748a652f04 ("Fix SIGBUS due to unaligned
access")
Also remove the existing patch file as the issue was resolved in the
mainline by 4c9a28a28e ("CMake: do not pass -march=native or
-mcpu=native when crosscompiling")
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Veselin Georgiev <veselin.georgiev@garmin.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
nlohmann-json has examples in their repo that refer to
the header as json.hpp directly, as such many packages
that require nlohmann-json expect json.hpp and not
nlohmann/json.hpp as the cmake file installs. Create symlink
so that access is available by either option. Also update
to release 3.3.0.
Signed-off-by: James Feist <james.feist@linux.intel.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Since `36983fe umip: move to meta-networking' applied,
it was moved to invalid location.
Fix prior partial move to meta-networking
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* devtool isn't able to regenerate them completely, it generates only
top level commit which says that submodules were changed, like this:
From c8fa114325e4e532657499b9bb6fd23cb6536eaa Mon Sep 17 00:00:00 2001
From: OpenEmbedded <oe.patch@oe>
Date: Tue, 2 Oct 2018 16:46:55 +0000
Subject: [PATCH] Committing changes from do_patch
---
external/libselinux | 2 +-
system/core | 2 +-
system/extras | 2 +-
3 files changed, 3 insertions(+), 3 deletions(-)
..
but it still creates relatively good commits in each submodule and
after small manual cleanup they can be exported with
git format-patch --no-numbered --no-signature
* rename .gitignore to prevent git ignoring the changes in .patch files
in meta-oe already (we want to use it only later during the build)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Ruslan Bilovol
Khem Raj
Ming Liu
Hong Liu
Alex Kiernan
Matthijs Vader
André Draszik
Mingli Yu
Peace Lee
Adrian Freihofer
Andrej Valek
Niko Mauno
Bartosz Golaszewski
Joshua Watt
Yi-Soo An
Tim Orling
James Feist
Hongxu Jia
Philip Balister
Martin Jansa