The issue is reported upstream:
https://bugs.ruby-lang.org/issues/18456
Otherwise I do not feel that further investigation of rdoc's parser to find
out the source of non-determinism is worth the time.
(From OE-Core rev: 208021f7212a8a790c350ccca720695c5bcbb1ca)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When the tranisiton to perl-cross occured, the threading define
seems to have been missed. The perl tests for threading where
simply skipped, so there was no direct failures. This was verified
by running perl ptest before and after the change to see PASS vs SKIP
results of threaded related tests.
NOTE: Perl officially discourges the use of threads, so this
functionality maybe depercated in the future [0][1]
v2: adds the usethreads to native and nativesdk. This was tested by
builing postresql and rrdtool which use perl and automake.
[0] https://perldoc.perl.org/5.34.0-RC2/threads#WARNING
[1] https://perldoc.perl.org/perlpolicy#discouraged
(From OE-Core rev: b9fd7cd319a1d8f0ddf5ea60710b015e9afb588c)
Signed-off-by: Saul Wold <saul.wold@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Without this, compiling gcc plugins will fail with an error similar to:
[...]
fatal error: ./config/i386/linux64.h: No such file or directory
[...]
In Yocto, we set up compiling gcc-cross out of tree. Which in turn makes
the generated headers end up in B. The tm.h header will include
generated headers that are expected in plugin/include/config/*.
For example, the linux64.h header, when generating gcc-cross for x86-64,
will end up in tm.h header as:
include "./config/i386/linux64.h"
On the other hand, the make rule `install-plugin` in gcc/Makefile.in
will install the linux64.h assuming that it is generated in the sources
directory and because this is not the case in our setup, the Makefile
ends up installing it in plugin_includedir/`basename $$path` which ends
up installing the header in [..]plugin/include as opposed to
[..]plugin/include/config/i386 (as expected by the generator of tm.h).
The included patch modifies the Makefile rule to match the assumption of
gcc-cross being compiled out of tree.
(From OE-Core rev: 92167f8e02bb6fbbe1ee6a6678525a0ae27b00a5)
Signed-off-by: Andrei Gherzan <andrei.gherzan@huawei.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Go has its own system for creating temporary build
sub-directories with randomized names, and setting
up debug-prefix-map on the fly to prevent those
directories leaking into target binaries. OE's own
settings were clashing with it, so this change
carefully avoids the two stepping on each other.
Additionally, the top level build directory cannot
be named 'go-something'.
(From OE-Core rev: 9985b17a30bb9b9f1bc82a44662687db5cead66e)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
go writes build-specific ids into binaries it produces
and has a custom system for calculating them from
file hashes, environment variables and other inputs
(not that dissimilar to sstate cache, actually). This can
go wrong :) in various ways (for purposes of reproducibility
in particular), so this enables useful logs to see what
happens and why.
(From OE-Core rev: a587be1d18fc55fe57d1aa5aa7c9e26af887109e)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
==========
Bugs fixed for this release <https://github.com/numpy/numpy/pull/#####>
#20357: MAINT: Do not forward __(deep)copy__ calls of _GenericAlias...
#20462: BUG: Fix float16 einsum fastpaths using wrong tempvar
#20463: BUG, DIST: Print os error message when the executable not exist
#20464: BLD: Verify the ability to compile C++ sources before initiating...
#20465: BUG: Force npymath to respect npy_longdouble
#20466: BUG: Fix failure to create aligned, empty structured dtype
#20467: ENH: provide a convenience function to replace npy_load_module
#20495: MAINT: update wheel to version that supports python3.10
#20497: BUG: Clear errors correctly in F2PY conversions
#20613: DEV: add a warningfilter to fix pytest workflow.
#20618: MAINT: Help boost::python libraries at least not crash
(From OE-Core rev: 956d6811ed0ae6db759a663866eed4977f994ad8)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In qemu, they merged the 32 and 64 bit riscv variants into one target for meson
purposes but this isn't officially supported by upstream meson. In normal builds
this would just be a warning but in OE, this is fatal.
Avoid the fatal errors for now and build as qemu intended whilst this is resolved
by the upstreams.
(From OE-Core rev: 7c1356e762d357dad95decf7b1d6147a2efae174)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
nativesdk-rpm build is also affected by the stray /var/tmp
created during the rpm install. Remove it to fix QA Issue:
nativesdk-rpm installs files in /usr/local/oe-sdk-hardcoded-buildpath/sysroots/x86_64-pokysdk-linux/var/volatile, but it is expected to be empty [empty-dirs]
(From OE-Core rev: 28dd1e1a0965646c5736e3de91aad830311a797b)
Signed-off-by: Samuli Piippo <samuli.piippo@qt.io>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There were warnings about non-existent files when there was
a duplicate in the remove-for-aarch list and the non-deterministic lists
so check if the files still exists before trying to deal with it.
Also restore the tests in the opposite order that they were hidden.
(From OE-Core rev: 2e1dc5782dcb879410d535f937c633edea484d11)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The drd/tests/boost_thread test passes when run interactively on qemuarm64
but fails for most runs when run in qemu on an arm worker such as ubuntu1804-arm-1
from the Yocto autobuilder cluster. Skip the test until we fix the problem.
(From OE-Core rev: 9a36983f9681f215aa2f59bfcb16343fe0a819cd)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The valgrind ptest script warns about tests such as:
mv: cannot stat 'drd/tests/bar_bad_xml.vgtest': No such file or directory
Since these tests were either in both remove-for-* lists or
were not installed since they fail everywhere as noted in:
85a0d61025 valgrind: fix the remaining ptest failures
984ffe3ab4 valgrind: Disable leak_cpp_interior test
remove them from the remove-for-aarch64 list
(From OE-Core rev: 9492cdc43b830ebf0d02ccaae342f33c826d8a01)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Using glibc-2.35-alpha, valgrind ptests mostly all fail with the error:
WARNING: unhandled amd64-linux syscall: 334
You may be able to write your own handler.
With this patch applied qemux86-64 ptest results were:
TOTAL: 728
PASSED: 709
FAILED: 0
SKIPPED: 19
This patch implements the Linux rseq syscall as ENOSYS while
a proper fix is being worked on as desribed in the patch
commit log.
(From OE-Core rev: 6ca008072df0371b6b1cfdd7030da6494ff012bd)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We are using taskset to improve the reproducability of tests listed in:
meta/recipes-devtools/valgrind/valgrind/taskset_nondeterministic_tests
so we need to have the ptests depend on util-linux-taskset.
(From OE-Core rev: aa06250cc4e258ad807eef9940939fd68c0bda8b)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We're seeing failures on systems with old selinux headers. Add a PACKAGECONFIG
entry so that we don't try and build this unless enabled. I'll leave
dependencies to someone who wants it and can test, I just need to fix
build failures right now as this shouldn't be enabled.
(From OE-Core rev: d26c86e67b1154b243d093ff80dafce54da47424)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This patch has been accepted upstream
(From OE-Core rev: 4c644708798f258d45dbcfa01909626de32710d4)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The test suite makes incorrect assumptions about OpenSSL versions post-
3.0, so backport the fix for the test suite.
(From OE-Core rev: 2b5e9cfd2c1914d606725029c59cd50cca3d5511)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add config option for pulseaudio
Change audio to default mode since quoting of list of devices is broken in meson
Drop patch where code changed completed upstream
Refresh other patches
(From OE-Core rev: c1685b06e5f3684b1b5631b99b412e3ec2b10800)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
which enables us to drop the -r option on qemu-static-riscv32
invocations.
(From OE-Core rev: 4552249810ed27d1272130226d054e7651a90ce6)
Signed-off-by: Matt Madison <matt@madison.systems>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The correct SPDX license for the test data contained in this code appears
to be Unicode-TOU. Update the LICENSE field to match and avoid package
QA warnings.
(From OE-Core rev: 7cd3f3649f478448db9f068e493cada03d04a85f)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
As that's what upstream prefers.
(From OE-Core rev: 5a6cd9cc1b9d8fd3607f3df311accb483d2989a3)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
================
Bugs fixed
----------
* Several incompatibilities with CPython 3.11 were resolved.
(Github issues #4411, #4414, #4415, #4416, #4420, #4428, #4473, #4479, #4480)
* Some C compiler warnings were resolved.
(Github issue #4439)
* C++ "std::move()" should only be used automatically in MSVC versions that support it.
(Github issue #4191)
* The "Py_hash_t" type failed to accept arbitrary "index" values.
(Github issue #2752)
* Avoid copying unaligned 16-bit values since some platforms require them to be aligned.
Use memcpy() instead to let the C compiler decide how to do it.
(Github issue #4343)
* Cython crashed on invalid truthiness tests on C++ types without "operator bool".
(Github issue #4348)
* The declaration of "PyUnicode_CompareWithASCIIString()" in "cpython.unicode" was incorrect.
(Github issue #4344)
(From OE-Core rev: 8d8faad71e05703362d4004cc9ecdb93c08398ec)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
