There's no need to have multiple overrides for this, as BBCLASSEXTEND
will automatically map the dependencies as needed.
(From OE-Core rev: ffa699332551fbbb95a0e388385667dc2706da6a)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This dependency has always existed in the oe-core recipes but isn't
explicitly needed here: the host Python is sufficient.
Note that rust-common still inherits on python3native[1] so this doesn't
yet actually have a meaningful change to the dependency tree.
[1] oe-core 4abd6ee9d4 ("rust-common.bbclass: rewrite toolchain wrappers in (native) python")
(From OE-Core rev: aee45fc067b2ccf3e365deb94584089b60cc7d4e)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop obsolete patch, the lib/gnulib.mk has been removed by upstream
- 0001-Unset-need_charset_alias-when-building-for-musl.patch
Drop backport patches:
- 0002-Fix-segfault-with-mangled-rename-patch.patch
- 0003-Allow-input-files-to-be-missing-for-ed-style-patches.patch
- 0004-Fix-arbitrary-command-execution-in-ed-style-patches-.patch
- 0001-Fix-swapping-fake-lines-in-pch_swap.patch
- CVE-2019-13636.patch
- 0001-Invoke-ed-directly-instead-of-using-the-shell.patch
- 0001-Don-t-leak-temporary-file-on-failed-ed-style-patch.patch
- 0001-Don-t-leak-temporary-file-on-failed-multi-file-ed.patch
- CVE-2019-20633.patch
GNU patch 2.8 released: http://savannah.gnu.org/news/?id=10741
NEWS since v2.7.6 (2018-02-03):
The --follow-symlinks option now applies to output files as well as input.
'patch' now supports file timestamps after 2038 even on traditional
GNU/Linux platforms where time_t defaults to 32 bits.
'patch' no longer creates files with names containing newlines,
as encouraged by POSIX.1-2024.
Patches can no longer contain NUL ('\0') bytes in diff directive lines.
These bytes would otherwise cause unpredictable behavior.
Patches can now contain sequences of spaces and tabs around line numbers
and in other places where POSIX requires support for these sequences.
--enable-gcc-warnings no longer uses expensive static checking.
Use --enable-gcc-warnings=expensive if you still want it.
Fix undefined or ill-defined behavior in unusual cases, such as very
large sizes, possible stack overflow, I/O errors, memory exhaustion,
races with other processes, and signals arriving at inopportune moments.
Remove old "Plan B" code, designed for machines with 16-bit pointers.
Assume C99 or later; previously it assumed C89 or later.
Port to current GCC, Autoconf, Gnulib, etc.
(From OE-Core rev: b7034d912122582bd63f06d2e4a849dd376b7157)
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It does not compile with clang due to include_next stdarg.h not working
as the system expects to match gcc behavior
(From OE-Core rev: 4ef959f37816f23e4ed57a71cb9a42fd818aa1fb)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Building of spirv-mesa and spirv64-mesa CLC targets (which are required
for Mesa to work) requires a working llvm-spirv tool (provided by the
spirv-llvm-translator-native package). Make clang build-depend on the
corresponding recipe in order to be able to build requried targets.
Fixes: 4178fe97371b ("clang: split SPIRV-LLVM-Translator to its own recipe")
(From OE-Core rev: 177aaa7912f317da4a17a57081eb4f5667ef2c02)
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The following bugs have been fixed or resolved in this point release:
503098 Incorrect NAN-boxing for float registers in RISC-V
503641 close_range syscalls started failing with 3.25.0
503914 mount syscall param filesystemtype may be NULL
504177 FILE DESCRIPTORS banner shows when closing some inherited fds
504265 FreeBSD: missing syscall wrappers for fchroot and setcred
504466 Double close causes SEGV
To see details of a given bug, visit
https://bugs.kde.org/show_bug.cgi?id=XXXXXX
where XXXXXX is the bug number as listed above.
(From OE-Core rev: 582e832634d5f1fa4ff9c89d095c10eaffcb3582)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
chrpath has limitations e.g. the original rpath in ELF have to bigger in size
than the one being edited into it by chrpath, some toolchains do not use RPATH
but emit the RUNPATHs into ELF files and chrpath is not able to handle the
runpaths, this is the case with mips and pp32 build of rust, especially
when using clang compiler to build them.
patchelf can do more:
Modify RUNPATH entries
Add RPATH/RUNPATH where none existed
Set longer paths than the original
Convert between RPATH and RUNPATH
(From OE-Core rev: 22b903f6620455e142e836412d3f7f6a4f03bea7)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
RISC-V multilib currently doesn't work at all. We could disable multilib for
the riscv platform but that would then behave differently to all our others.
Instead, copy the non-multilib config over the multilib config for now for
riscv, meaning we can keep the platforms similar.
This isn't quite enough as the triplet specific c++ headers are in the wrong
place leading to compiler issues and testimage failures. Work around that too
until someone adds full multilib support for the platform (if desired).
(From OE-Core rev: 3081f62c18fcee642ab43efa717c8f71d51ae587)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Some unit tests were previously marked as ignored due to failures in earlier
versions of Rust. With the upgrade to Rust 1.85.1, these tests are now passing
consistently. They've been re-enabled and verified to run successfully on the
latest version.
(From OE-Core rev: 56149c8c2108973666251e21609a7210a91984cf)
Signed-off-by: Deepesh Varatharajan <Deepesh.Varatharajan@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In rpm-sequoia, some default paths in test code (OUT_DIR,
CARGO_MANIFEST_DIR) are invalid at runtime and cause test failures.
To fix this, patch the test code (symbols.rs) to support overriding these
paths via optional environment variables: FORCE_RUNTIME_PATH_LIB and
FORCE_RUNTIME_PATH_SRC.
Also make -ptest package RDEPEND on -dev package.
Tests take less than a second so this is added to PTEST_FAST.
ptest result:
|root@qemux86-64:~# ptest-runner rpm-sequoia
|START: ptest-runner
|2025-05-02T15:57
|BEGIN: /usr/lib/rpm-sequoia/ptest
|
|running 1 test
|test symbols ... ok
|
|test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s
|
|
|running 2 tests
|test tests::merge_certs_mismatch ... ok
|test tests::merge_certs ... ok
|
|test result: ok. 2 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.01s
|
|DURATION: 0
|END: /usr/lib/rpm-sequoia/ptest
|2025-05-02T15:57
|STOP: ptest-runner
|TOTAL: 1 FAIL: 0
(From OE-Core rev: 16499cf903718e6bf022a13baa09df610cd43d62)
Signed-off-by: Ines KCHELFI <ines.kchelfi@smile.fr>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Many changes are included with this release, including support for
freethreading builds of CPython 3.13, cleanup of legacy code, and many
other features and bug fixes.
Changelog: https://github.com/cython/cython/blob/master/CHANGES.rst
Patch '0001-Output-import-relative-paths-in-generated-C-code.-GH.patch'
is included in 3.1.0, so we no longer need the backport:
|tgamblin@megalith ~/workspace/git/pythonsrc/cython (master)$ git tag --contains 20bceea6b19ffc2f65b9fba2e4f737f09e5a2b20
|3.1.0
|3.1.0-1
|3.1.0a1
|3.1.0b1
|3.1.0rc1
|3.1.0rc2
Reproducibility looks OK.
(From OE-Core rev: 26a73392524f648015d55bf421a9b1bf5ac0d955)
Signed-off-by: Trevor Gamblin <tgamblin@baylibre.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add logic to deduce compiler-rt target architecture based on tune
features
Cleanup the cmake options for native/nativesdk/target recipes
(From OE-Core rev: 579d4d85574624e39140fb74f6d37dcd7bf363a6)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Using target triple confuses the native system compiler to find
libgcc and C runtime during compiler-rt-native build
Fixes
| /home/khem/yoe/build/tmp/hosttools/ld: cannot find crtbeginS.o: No such file or directory
| /home/khem/yoe/build/tmp/hosttools/ld: cannot find -lgcc: No such file or directory
| /home/khem/yoe/build/tmp/hosttools/ld: cannot find -lgcc_s: No such file or directory
| clang: error: linker command failed with exit code 1 (use -v to see invocation)
| ninja: build stopped: subcommand failed.
(From OE-Core rev: 1d4676d6b7c0fe40ab4f808dda436dc2f2915fd8)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
arm architecture < v6 does not have the dmb support, therefore do
not add the sync primitives to compiler-rt builtins build
(From OE-Core rev: e1b143860b221600f7dab5c08e5b187aefb851dd)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
=============
- Rule comments: tell how many spaces are expected
- Rule quoted-strings: Fix only-when-needed on multiline with backslash
- Config: Report if rules is not a dict
- Fix test_codec_built_in_equivalent() test when run with pytest
- CI: Fix TestPyPI "dev0" versions for master commits on tags
- Docs: Add links to GitHub repository and releases
- Docs: Fix GitLab integration example
- Docs: Fix GitLab integration link
- Fix the tests badge link on the README
(From OE-Core rev: b49426b3577546e197c42aa28b5af4a0f3ea1c53)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
=============
- Fix TypeError when taking the union of typing_extensions.TypeAliasType and a
typing.TypeAliasType on Python 3.12 and 3.13.
- Backport from CPython PR #132160 to avoid having user arguments shadowed in
generated __new__ by @typing_extensions.deprecated.
(From OE-Core rev: f01c7ece71f4d9887763ee5062c56d454f88ae3d)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changed
----------
- Update list of supported licenses
Fixed
--------
- Fix an issue where the union of specific inverse or partially inverse markers
was not simplified
- Fix an issue where optional dependencies defined in the project section were
treated as non-optional when a source was defined for them in the tool.poetry
section
- Fix an issue where markers with === were not parsed correctly
- Fix an issue where local versions with upper case letters caused an error
- Fix an issue where extra markers with a value starting with "in" were not
validated correctly
- Fix an issue where inheriting from WheelBuilder was unnecessarily difficult
(From OE-Core rev: ca51448b82abd9333ad2763d52c473cbe876d5c5)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Features & Improvements
------------------------
- New command pdm new that behaves like pdm init but creates a new project.
- Support use --name as project name for command pdm new e.g. pdm new hello --name world
- Support exporting to pylock.toml format as described by PEP 751.
Bug Fixes
---------
- Pass the --quiet option to pdm sync command.
- If a .python-version file is found and it contains multiple lines, the file
will be ignored. The usage of the .python-version file can be disabled, if
configuration value python.use_python_version (or environment variable PDM_USE_PYTHON_VERSION) is False.
- fix pdm config -e command to open read-only file under linux
- Replace project names and import names in both README.md and pyproject.toml when running pdm init <template>.
- Fix a bug that URL dependency hashes are not updated if running pdm lock --update-reuse.
- Install the project when using the BaseSynchronizer with install_self set
to True. This fixes the bug that when calling pdm sync --quiet, it skips
installing the project itself.
- Mark one additional test as requiring network, and fix another one
not to require it anymore.
(From OE-Core rev: 01cacd208c7a3e1fa2923ef346d7a1c65a4005c0)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
============
- Re-add a test for Unicode file name parsing
- Upgrade to ruff 0.9.1
- Add support for PEP 738 Android tags
- feat(markers): support 'extras' and 'dependency_groups' markers
(From OE-Core rev: 1f6a72bf37297a362119375523750544a11a23ea)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
===========
Changed
-------
- DRY fix in abbr extension by introducing method create_element
- Clean up test directory by removing some redundant tests and port non-redundant cases to the newer test framework.
- Improved performance of the raw HTML post-processor
Fixed
---------
- Backslash Unescape IDs set via attr_list on toc
- Ensure md_in_html processes content inside "markdown" blocks as they are parsed outside of "markdown" blocks to keep things more consistent for third-party extensions
- md_in_html handle tags within inline code blocks better
- md_in_html fix handling of one-liner block HTML handling
- Ensure <center> is treated like a block-level element
- Ensure that abbr extension respects AtomicString and does not process perceived abbreviations in these strings
- Ensure smarty extension correctly renders nested closing quotes
(From OE-Core rev: 662d586edb3afed8273ec4910ea1a4c090f8b757)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
- Fix undefined variable errors when strict_undefined=True when using a
nested list comprehension.
(From OE-Core rev: 40a60f0ca45116604430f8b0d1ee4f70e1f9843e)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bugfix:
-Binary wheels use libxml2 2.13.8 and libxslt 1.1.43 to resolve several CVEs.
(Binary wheels for Windows continue to use a patched libxml2 2.11.9 and libxslt 1.1.39.)
(From OE-Core rev: 4e15eededc4c67665c48c0fcdcfa41cfd0d3bf40)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
==========
- $HFPR_PACKAGE_NAME is now replaced by the package name in the PyPI readme.
- Support for Python 3.7.
(From OE-Core rev: 5d3f3f8f9770e81bd7c2d53a1512577792ba4fa2)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
============
- Added the 's' specifier for 'std::error_code'. It allows formatting an error
message as a string.
- Fixed formatting of 'std::chrono::local_time' and 'tm'
- Added diagnostics for cases when timezone information is not available.
- Deprecated 'fmt::localtime' in favor of 'std::localtime'.
- Fixed compilation with GCC 15 and C++20 modules enabled
- Fixed handling of named arguments in format specs
- Added error reporting for duplicate named arguments
- Fixed formatting of 'long' with 'FMT_BUILTIN_TYPES=0'
- Optimized 'text_style' using bit packing
- Added support for incomplete types
- Fixed a flush issue in 'fmt::print' when using libstdc++
- Fixed 'fmt::println' usage with 'FMT_ENFORCE_COMPILE_STRING' and legacy
- Removed legacy header 'fmt/core.h' from docs
- Worked around limitations of '__builtin_strlen' during constant evaluation
- Worked around a bug in MSVC v141
- Removed the 'fmt_detail' namespace
- Removed specializations of 'std::is_floating_point' in tests
- Fixed a CMake error when setting 'CMAKE_MODULE_PATH' in the pedantic mode
- Updated the Bazel config
(From OE-Core rev: 2f2ef80e7e037900d8d4afa3dc7bf01734cef229)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
License-Update: Copyright year updated to 2025
(From OE-Core rev: e96c1a869d250250f14ff2f0089df78b7ec5795c)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently, pseudo tracks all files referenced within its presence unless
they're listed in an exclusion list. The exclusion list has grown to be
fairly unwieldy.
This patch swaps PSEUDO_IGNORE_PATHS for PSEUDO_INCLUDE_PATHS which in
theory should be easier and more explicit to maintain.
This change does drop many directories from pseudo coverage including
/home and /tmp. There may be adapatations needed for recipes/classes
using pseudo in specific ways.
(From OE-Core rev: 2502da81709f25de499277b28d33c915638c45f6)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
OE does not rely on tuple to deduce hardfloat ABI, but clang/llvm does
arm-yoe-linux-gnueabi is used for both soft and softfp and hardfp float
ABIs in OE, LLVM expects arm-yoe-linux-gnueabihf for it to be treated as
hardfloat ABI, and look for correct name for rt libraries.
We know when we compile them with Hard-float ABI so rename them in such
case so clang can find it when using -rtlib=compiler-rt it needs to
has 'armhf' suffix
(From OE-Core rev: b369e99cb27bd327ab244335b637f0ad3393d9b8)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
-stdlib is needed in both cxxflags and ldflags since c++ compiler
will use these to chose the right c++ headers and libraries during link
-rtlib and -unwindlib are only needed during linking
Use LLVM_HOST_TRIPLE during cross compiling compiler-rt for target
this is used by cmake to compute target and compiler triples
Fix LLVM_RUNTIME_TARGETS to use HOST_ARCH, it does not need cross
compile triple which HOST_SYS is passing here
(From OE-Core rev: d1ae4087c548a78e90687b64764f621b070e3ec5)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Specify CMAKE_C_COMPILER_TARGET and COMPILER_RT_DEFAULT_TARGET_ARCH
in common for native/target/nativesdk
set COMPILER_RT_DEFAULT_TARGET_ARCH to use HOST_ARCH across all recipe
variants
It is not needed to set COMPILER_RT_DEFAULT_TARGET_ARCH for powerpc anymore
as it is already fixed by setting COMPILER_RT_DEFAULT_TARGET_ARCH to
HOST_ARCH
(From OE-Core rev: 945e9a799ddbb28c169155368f1998cb8e9f455f)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
