Set GRPC_PYTHON_BUILD_EXT_COMPILER_JOBS to limit spawned compiler
processes. Without this it uses all available CPUs (via
multiprocessing.cpu_count()) and can exhaust build host since there are
lot of files to compile (e.g. with 128 cores it manages to spawn 128 gcc
processes)
Note that this is a general problem for all setuptools based builds with
build_ext compilation which can either compile with 1 thread or
cpu_count threads. grpcio hot-patches setuptools and allows to set
specific build concurrency value.
Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
disable_libunwind_aarch64.patch is dropped because from my runtime
testing, this patch is no longer needed.
0001-disbale-heap-checkers-and-debug-allocator-on-musl.patch is
dropped because I don't see it necessary now. Things still build
without it.
0001-src-mmap_hook.cc-Fix-build-for-32bit-machine.patch is dropped
because this oe specific patch is no longer needed because of the
following two commits:
02adc8c also disable _TIME_BITS in mmap_hook.cc
198b3dd disable _FILE_OFFSET_BITS in mmap_hook.cc
ppc-musl.patch is rebased.
0002-src-base-elf_mem_image.cc-fix-build-for-musl.patch is added
to fix build failure for musl + ppc64.
0003-Makefile.am-disable-building-noinst-tests-for-musl.patch is
added as a workaround to fix build failure on musl + ppc64.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
As of right now there are no applications available for benchmarking
OpenCL. Clpeak is a simple synthetic benchmark for OpenCL that is
designed to determine the peak capabilities of OpenCL devices.
It's currently one of the more popular CL benchmarks supported by
OpenBenchmarking.org and licensed Apache-2.0.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This is a backport of a fix to iperf3. The author saw a 40% improvement in
their network throughput, we've seen around a 55% improvement in our tests.
Link: ac6b9f7fd3
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Recent change to autotools.bbclass broke netperf as it can not now find
proper macros. Remove old workaround and add a proper patch to utilize
ACLOCAL_AMFLAGS option.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The liberation font release 0.2 is not available for download
anymore and was released in 2007, so let's do an update.
Since 2007 also the license has changed to a better suited
SIL Open Font License. In addition, the ttf is now build from
source like, e.g., in ttf-lohit.
Signed-off-by: Thomas Schlien <ts@ferncast.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
BUILD_CFLAGS doesn't contain information about the recipe sysroot nor about the target
system - and in case these are not specified for bindgen, then it builds against
the build system's headers, using the build system's arch.
Instead of BUILD_CFLAGS pass HOST_CC_ARCH and TOOLCHAIN_OPTIONS, which contain the sysroot
and the requred cross-compiling flags, beside passing the target flag also.
Fixes the following do_compile error:
| /usr/include/linux/types.h:5:10: fatal error: 'asm/types.h' file not found
| thread 'main' panicked at /yocto/sandbox/build/tmp/work/cortexa72-poky-linux/thin-provisioning-tools/1.1.0/sources-unpack/cargo_home/bitbake/devicemapper-sys-0.3.0/build.rs:24:10:
| Could not generate dm.h bindings: ClangDiagnostic("/usr/include/linux/types.h:5:10: fatal error: 'asm/types.h' file not found\n")
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Add a patch to not use AM_INIT_AUTOMAKE as automake isn't actually used,
and just let autotools.bbclass run autoreconf.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Add some implicit dependencies which are explicitly checked for, a small
comment, and remove an obsolete FILES:PN-dbg.
Signed-off-by: Ross Burton <ross.burton@arm.com>
When unpacking, just use cp to hardlink the xserver source tree into the
tigervnc source tree.
When configuring, instead of replicating complicated logic, we can
simply set the variable AUTOTOOLS_SCRIPT_PATH and just call
autotools_do_configure. Some of the options passed to the xserver
configure are obsolete so remove them.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Pass the correct path so the unit is installed where it should be, and
use the systemd class to ensure the unit is enabled correctly.
Signed-off-by: Ross Burton <ross.burton@arm.com>
The configure scripts do the right thing now, so there's no need to
patch configure.ac or manually invoke the sub-configure.
Also remove obsolete FILES, -dbg is packaged automatically and we don't
install .la files.
Signed-off-by: Ross Burton <ross.burton@arm.com>
As of "autotools: don't try and find in-tree macros" in core, acpaths is
no longer used.
Also switch to out-of-tree builds as this appear to work now.
Signed-off-by: Ross Burton <ross.burton@arm.com>
As of "autotools: don't try and find in-tree macros" in core, acpaths is
no longer used.
Also remove an obsolete FILES:PN-dbg as the debug package is packaged
automatically.
Signed-off-by: Ross Burton <ross.burton@arm.com>
No need to manually write do_configure, the config.h.in is generated
with autoheader so don't disable that, but disable aclocal as that file
is hand-maintained.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Iksemel is a XML parser library that was designed for Jabber applications,
but it has no users and hasn't been maintained for 14 years.
Signed-off-by: Ross Burton <ross.burton@arm.com>
If zsh is built with the api-documentation enabled, the manpages.bbclass
adds the 'manpages' PACKAGECONFIG but this isn't defined:
ERROR: zsh-5.9-r0 do_configure: QA Issue: zsh: invalid PACKAGECONFIG: manpages [invalid-packageconfig]
zsh doesn't have an option to disable the manpages, so add a stub
PACKAGECONFIG to silence this error.
Signed-off-by: Ross Burton <ross.burton@arm.com>
cpuid's Makefile passes -s to install to strip the binaries, but as this
uses the _host_ strip it can fail if that strip doesn't know about the
target (for example, building cpuid for x86 on aarch64):
install -D -s -m 755 cpuid .../image/usr/bin/cpuid
strip: Unable to recognise the format of the input file `.../image/usr/bin/cpuid'
Set INSTALL_STRIP='' so that the binaries are not stripped, and remove
the INSANE_SKIP for already-stripped.
Signed-off-by: Ross Burton <ross.burton@arm.com>
This is a major version upgrade, see release note:
https://www.postgresql.org/docs/release/
* Refresh patches:
0003-configure.ac-bypass-autoconf-2.69-version-check.patch
0005-postgresql-fix-ptest-failure-of-sysviews.patch
* Removed dropped option --enable-thread-safety
* Add depends for bison-native
* Fix do_packge_qa error:
*.c *.h generated by bison or flex leave full paths in comment,
rewrite those before *-src packaging
Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The API documentation should be buildable with gtk-doc, but it fails
for some reason:
gtkdoc-mkhtml: error: unrecognized arguments: ../gstd-docs.xml
Leave the gtk-doc class inherited, but temporarily disable it until this
is fixed.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
autoreconf needs to be told where to find macros as the Makefile.am does
not do this.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The manpages.bbclass adds the 'manpages' PACKAGECONFIG but this isn't
defined:
ERROR: QA Issue: udpcast: invalid PACKAGECONFIG: manpages [invalid-packageconfig]
There is not an option to disable the manpages, so add a stub
PACKAGECONFIG to silence this error.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This recipe doesn't ship any manpages since 2.10, so remove the manpage
support.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Set correct pkgconfig directory via INSTALL_PCDIR, otherwise on 64-bit
targets, the .pc file will be installed to /usr/lib64/pkgconfig by
default.
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Set pam module path to ${base_libdir}/security via INSTALL_PAMDIR. Then
we can get rid of the workaround in do_install.
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
License-Update:
change LICENSE to LICENSE.txt
Copyright year updated to 2025.
fix license to follow SPDX format
Changelog:
===========
- remove non-Linux support from all compilation paths (OS_KERNEL_NAME)
- add an aarch64 cross-compilation (requested by DAOS)
- mute error messages when transactions are intentionally aborted
- mute error message "Cannot find any matching device, no bad blocks found" when PMDK is used without PMem
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
===========
- tests: Fix some typos in the ipmisol basic test
- tests: Fix the modemstate and add linestate tests
- Fix modemstate and linestate handling
- Fix break sending to use the right function
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
==========
- Hold a strong reference to the AsyncEngine setup task
- docs: Enable link to source code
- Non unique name during wheel upload
- Add a helpful hint for when EADDRINUSE happens during startup
- Wheel builds failing after adding armv7l builds
- Add armv7l wheel builds
- Make no buffer space available when adding multicast memberships forgiving
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
===========
- Now supports Python 3.9 to 3.13, and no longer requires backports.zoneinfo.
- Debian is for some reason removing support for /etc/timezone, which is bad,
because that's the only place where the timezone is stated in plain text,
and what's worse, they don't delete it. So we can't trust it now,
so when we have multiple configs, we are forced to just ignore it.
- Attempts to return a ZoneInfo object also for UTC.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
=============
- GZipMiddleware: Make sure Vary header is always added if a response can be compressed
- Raise exception from background task on BaseHTTPMiddleware
- GZipMiddleware: Don't compress on server sent events
- MultiPartParser: Rename max_file_size to spool_max_size
- Add deprecated warning to TestClient(timeout=...)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
