Full release notes: https://valgrind.org/docs/manual/dist.news.html
In summary, there is a new configure option: --with-gdbscripts-dir
that lets you install the gdb valgrind python monitor scripts in a specific location.
It's not used in the valgrind recipe yet. Also, there were a few Memcheck and Cachegrind
improvements made and many bugs fixed.
The update required removing the patch:
0001-none-tests-x86-linux-seg_override.c-add-missing-incl.patch
as it's dealt with by these commits:
f7e4bb4af Bug 382034 - Testcases build fixes for musl
306232d40 musl: another fix for building testcases
as described on: https://bugs.kde.org/show_bug.cgi?id=382034
Drop: 002-context-APIs-are-not-available-on-musl.patch
since this is handled in a different way by the following
upstream commits:
c9e88f345 configure, drd: Only build the swapcontext test if swapcontext() is available
7cd4d7816 memcheck/tests/linux/stack_changes: Only run this test if setcontext() is available
And the that tracked the error:
https://bugs.kde.org/show_bug.cgi?id=434775
is closed.
No regression in musl qemux86-64/kvm ptest results.
Drop: 0001-fix-opcode-not-supported-on-mips32-linux.patch
The resolved (works for me) upstream defect rejected this patch:
https://bugs.kde.org/show_bug.cgi?id=396905
and suggested using CFLAGS="-mips32". With this patch dropped, the
qemumips build succeeds perhaps due to using -march=mips32r2
Drop: 0001-Make-local-functions-static-to-avoid-assembler-error.patch
since this was fixed upstream by commit:
d6da48fe5 mips: use local labels for do_acasW()
Confirmed with:
MACHINE=qemumips TCLIBC=musl bitbake valgrind
Ptest Resutls for qemux86-64/kvm
glibc:
=== Test Summary ===
TOTAL: 792
PASSED: 759
FAILED: 14
SKIPPED: 19
musl:
=== Test Summary ===
TOTAL: 792
PASSED: 559
FAILED: 211
SKIPPED: 22
(From OE-Core rev: a04c7714e9f69b7a3fc36adecf7e55c0fcfaa891)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When the helper scripts were split into separate packages, they
weren't added as ptest dependencies. Fix that.
(From OE-Core rev: 81c15142889ccb6cc2f2a00e7cdb92bcb36ca12c)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Specifically this fixes the issue where if a pytest test suite fails
during collection then the errors are hidden.
(From OE-Core rev: db0e82135ce73d0d6d55b2c2ac17a3fdec8aca99)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This fix some host contaminations issues:
| /poky/build/tmp/hosttools/ld.bfd: libcommon.fa.p/net_af-xdp.c.o: undefined reference to symbol 'bpf_xdp_detach@@LIBBPF_0.7.0'
| /poky/build/tmp/hosttools/ld.bfd: /usr/lib/libbpf.so.1: error adding symbols: DSO missing from command line
| collect2: error: ld returned 1 exit status
The AF_XDP network backend support [1] requires the libxdp that is not
available in any layer, otherwise you can configure this option:
| PACKAGECONFIG[af-xdp] = "--enable-af-xdp,--disable-af-xdp,libxdp"
[1] cb039ef3d9
(From OE-Core rev: c2524614e49aee95b23d7bc162b1a30dee8cf59c)
Signed-off-by: Jose Quaresma <jose.quaresma@foundries.io>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For some reason dataclasses was being packaged in the -profile subpackage,
which doesn't make sense as this is a core piece of the runtime support.
(From OE-Core rev: bc17d33b30a6d5d970f390becf726e852fd0cc16)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
-Added OpenSSL.SSL.Connection.get_selected_srtp_profile to determine which SRTP
profile was negotiated.
(From OE-Core rev: fc93e8cfe0dde7dd17bd55bccc8a64011b26f1b1)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
============
-A warning :class:'~pluggy.PluggyTeardownRaisedWarning' is now issued when an
old-style hookwrapper raises an exception during teardown. See the warning
documentation for more details.
-Add :func:'PluginManager.unblock <pluggy.PluginManager.unblock>' method to
unblock a plugin by plugin name.
-Fix :func:'~pluggy.HookCaller.call_extra()' extra methods getting ordered
before everything else in some circumstances. Regressed in pluggy 1.1.0.
-Fix plugins registering other plugins in a hook when the other plugins
implement the same hook itself. Regressed in pluggy 1.1.0.
(From OE-Core rev: 1bc326d5228e90bd3cd19d7f3f10b2f74a39d06e)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
Fix loading of local plugins to account for newly released versions of a dependency
(From OE-Core rev: f372bcf6d530bb0627fe2ed49e52e7ae279bf77f)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bugfix:
---------
Fixed parsing issue where attempting to render a single percent sign %
using an escaped percent %% would not function correctly if the escaped
percent were not the first character on a line.
License-Update: Copyright year updated to 2024.
(From OE-Core rev: 6cc0e19da1259b4461e6b9c7014ab041b9d60fa4)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
License-Update: Copyright year updated to 2024.
(From OE-Core rev: 3752a33ce9490d8fc9d59a0e952faa47c0abba71)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
===========
- Remove Sphinx as a required dependency, as circular dependencies may cause
failure with package managers that expect a directed acyclic graph (DAG)
of dependencies.
(From OE-Core rev: d82e4f1f571e90adc2290eb819f4c8b6ebaa3d1a)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Upstream is in the process of migrating testing to pytest, backport
test_dump_load.py (the legacy_tests were throwing Errors and Failures).
We don't really need to run upstream's test suite, but we do need to verify
that our runtime is functional.
a98fd6088e
(From OE-Core rev: b0fdf0688109ba6b87840a0837bebee3d9a27089)
Signed-off-by: Tim Orling <tim.orling@konsulko.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For some time now, we have probably been susceptible to host contamination
or at least non-deterministic behavior when libyaml was available in the
build environment.
The symptom is:
ERROR: python3-pyyaml-6.0-r0 do_package_qa: QA Issue:
/usr/lib/python3.11/site-packages/yaml/_yaml.cpython-311-aarch64-linux-gnu.so
contained in package python3-pyyaml requires libyaml-0.so.2()(64bit), but no
providers found in RDEPENDS:python3-pyyaml? [file-rdeps]
>From the documentation:
"""
By default, the setup.py script checks whether LibYAML is installed and
if so, builds and installs LibYAML bindings. To skip the check and force
installation of LibYAML bindings, use the option --with-libyaml:
python setup.py --with-libyaml install
To disable the check and skip building and installing LibYAML bindings,
use --without-libyaml:
python setup.py --without-libyaml install
"""
Instead of leaving this to chance, add PACKAGECONFIG and by default build
with the faster libyaml bindings.
(From OE-Core rev: dfde9526f9183907b2bc47fde4f59ab3a5848d90)
Signed-off-by: Tim Orling <tim.orling@konsulko.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The opkg-utils HOMEPAGE value should follow the yocto opkg fork - which
has been moved off of Google Code and is now centralized on yocto git
infrastructure.
(From OE-Core rev: 6ff8b43e08840582ee8d88261cbaffd33c2304f9)
Signed-off-by: Alex Stewart <alex.stewart@ni.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The opkg internal sat-solver is marked for deprecation in a future
release.
Add a recipe QA check to warn users who have the internal solver
configured.
(From OE-Core rev: c8405b86c6c3f9054810914105de5d70c4b7800b)
Signed-off-by: Alex Stewart <alex.stewart@ni.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When debugging a C++ program on target using libcstdc++, gdb does not
find the pretty printer python support, this is because the init file is
installed into /usr/lib which is excluded from gdb's default list of
directories to load config's from, where as defaults to load from
datadir is still on, it therefore does not load this file
warning: File "/usr/lib/libstdc++.so.6.0.33-gdb.py" auto-loading has been declined by your `auto-load safe-path' set to "$debugdir:$datadir/auto-load".
This leaves the libstdc++ pretty-printers disabled even though added via
IMAGE_INSTALL += "gcc-runtime-dbg"
Comment on adding missing runtime dependency on datetime python module for the pretty
printer to work/load successfully.
Once the pretty-printer gdb file is moved out of libdir, it may become
empty therefore delete it if its empty.
(From OE-Core rev: d24e16d192f21aa68b5274496efaba1d913f37c2)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
syslinux has vendored copy of ext2fs/ext2_fs.h but uses ext2fs/ext2fs.h
from e2fsprogs package, however, ext2fs/ext2fs.h has dependencies on
ext2fs/ext2_fs.h coming from e2fsprogs package as these both headers
come from same package, here syslinux uses ext2fs.h from e2fsprogs but
supplies its own copy of ext2_fs.h which maybe out of sync and that
results in warnings about implicit implicit-function-declarations
e.g.
recipe-sysroot/usr/include/ext2fs/ext2fs.h:727:16: error: implicit declaration of function 'ext2fs_has_feature_gdt_csum' [-Wimplicit-function-declaration]
| 727 | ext2fs_has_feature_gdt_csum(fs->super);
| | ^~~~~~~~~~~~~~~~~~~~~~~~~~~
ext2fs_has_feature_gdt_csum here comes from newer version of
ext2fs/ext2_fs.h but missing from vendored copy, hence the warning.
With gcc-14 this warning is treated as error by default, which breaks
the build, so lets treat it as warning only.
All these functions are never used in syslinux, so functionality-wise we
are fine.
(From OE-Core rev: a2b30108055e68b62fdad7319d7d569bc38a07b4)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This was a bugfix release containing several important fixes according to upstream.
Drop a backported patch.
(From OE-Core rev: 1715d0907d372e787c054a5509690a2719acfed0)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This helps working with native qemu recipes with devtool, since devtool
inserts an internal class called devtool-source resulting in
ERROR: QA Issue: qemu-system-native: native/nativesdk class is not inherited last, this can result in unexpected behaviour. Classes i
nherited after native/nativesdk: devtool-source.bbclass [native-last]
ERROR: Fatal QA errors were found, failing task.
ERROR: Parsing halted due to errors, see error messages above
ERROR: Command execution failed:
(From OE-Core rev: 6754387d7bd397700cef029ae6df86ff8cb3a4af)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Move away from the outdated setuptools to more modern build
classes.
(From OE-Core rev: dd0de1f1dcbbc389ceb1a032c393e5148e01ad98)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Note that 2.1.4 broke python3-jinja2 so skip to 2.1.5 which doesn't.
(From OE-Core rev: 697a249f27d16adeeee074d0baad83ad67c7a9e6)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* Drop CheckLib.pm patch; no longer applies
* Change to CPAN_MIRROR for SRC_URI
* Drop SRC_URI[md5sum]
* Drop redundant ptest-perl/run-ptest; ptest-perl inherit does this
* Drop sed for Expat/Makefile; no longer shipped
* Inherit pkgconfig to find expat lib
* Patch Makefile.PL to not try to load host expat
License-Update: Use LICENSE file; change to Artistic-2.0
https://metacpan.org/dist/XML-Parser/changes
2.47 2023-12-28 (by Todd Rinaldo)
- #84 use $fh instead of $foo
- #85 Fix typo in documentation
- #89 Devel::CheckLib to from 0.99 -> 1.14
- Devel::CheckLibn 1.16
- #91 POD fix for verbatim text
- #97 Add a LICENSE file
- #94 Don't ship Expat/Makefile
- Various github workflow improvements. Windows is still not working.
(From OE-Core rev: 5c51810772a6611a37b2c1b0a32d9558007c4cf7)
Signed-off-by: Tim Orling <tim.orling@konsulko.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
===========
* Do not flatten the solver solution into the original world used as
input to the solver.
* Fix warnings with GCC 14 -Walloc-size.
* Add --solution to the pkgconf CLI to dump the solver state.
* Improve the --digraph output to clarify cancelled edges in a given
solution.
* Demote requires dependencies to requires.private when a parent
dependency is pulled in via requires.private.
* Trim trailing whitespace when processing package arguments.
* Avoid strncmp() in --modversion version comparison.
* Update autoconf compile flag checking macro.
* Add system default path configuration to Meson.
* Fix order of PKG_CONFIG_LIBDIR and PKG_CONFIG_PATH element processing.
(From OE-Core rev: eacbec50fab40dd5c6b16b018806800526dfbfab)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
==============
- cache scheme so it never attempt to load it again
- Really revert "use Scalar::Util::reftype instead of ref to check for
ARRAY"
- Revert the reftype change introduced in 5.22 as it causes warnings.
- Use Scalar::Util::reftype instead of ref to check for ARRAY
(From OE-Core rev: 84fb221f02abb8e34a3fc8efab7c2725e6c1d144)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
============
- add Test2 compatibility in the test triggered by done_testing();
- new long-awaited feature (!!!) "allow_patterns", to either
temporarily or globally exempt particular warning patterns
(From OE-Core rev: ed3940f403ea4b9b2d875595953863dd28d94b8b)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
rpm 4.19 now builds with LFS64 support enabled by default,
so it calls statvfs64() to get the space available on the
filesystem it is installing packages into. This is not
getting caught by pseudo, so rpm is checking the host's
root filesystem, rather than the filesystem where the
build is happening.
Merge in that fix and a gcc14 fix.
(From OE-Core rev: f6d021c860b2b99f46c604149317b326f493022d)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The dependency `ima-evm-utils` of package config `imaevm` has been
removed during rpm upgrade. Add it back, otherwise it fails to run
do_configure when the package config `imaevm` enabled.
(From OE-Core rev: 15df110b79bfa2428f70453aa6aabe7239f8e25a)
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
