Peter Marko [Wed, 27 Nov 2024 19:23:33 +0000 (20:23 +0100)]
init-manager: define dev_manager for all init managers
Currently VIRTUAL-RUNTIME_dev_manager is defined by some init managers
with "??=" and in packagegroup-core-boot with "?=".
This means that this variable is different in this package group and in
all other the recipes.
This was discovered when trying to use new feature INIT_MANAGER when
migrating to scarthgap and using systemd distro. After deleting all
VIRTUAL_RUNTIME providers defined in init-manager-systemd udev was
installed additionally via packagegroup-core-boot.
Having a distro settings overriden in single recipe is wrong and needs
to be corrected. Therefore let's define the setting in all init managers
and remove it from packagegroup-core-boot.
core-image-tiny-initramfs has a dilemma - use busybox-mdev even if
distro does not enable it in busybox recipe to keep the image tiny, or
rather not install it even if distro enables it in busybox.
This patch chooses the first option.
Signed-off-by: Peter Marko <peter.marko@siemens.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Jamin Lin [Wed, 20 Nov 2024 06:17:24 +0000 (14:17 +0800)]
kernel-uboot: Support lzma compressed algorithm
Linux kernel supports to compress its output binary with lzma algorithm and
U-Boot support to decompress lzma compressed image.
Both xz and lzma algorithms are identical and users can use xz utility with
format parameter to compress Image which supports lzma algorithm.
Add the following command to support lzma. This command is refer to Linux
kernel, https://github.com/torvalds/linux/blob/master/arch/powerpc/boot/wrapper
"xz --format=lzma -f -6 linux.bin"
Signed-off-by: Jamin Lin <jamin_lin@aspeedtech.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Thomas Perrot [Tue, 26 Nov 2024 14:35:58 +0000 (15:35 +0100)]
tcl: resolve conflicts with tcl8's man pages
Change the suffix from 'tcl' to 'tcl9' and from 'tcl8' to avoid
conflicts with the man pages provided by the tcl packages.
file /usr/share/man/mann/vwait.ntcl conflicts between attempted installs of tcl-doc-9.0.0-r0.riscv64 and tcl8-doc-8.6.15-r0.riscv64
file /usr/share/man/mann/while.ntcl conflicts between attempted installs of tcl-doc-9.0.0-r0.riscv64 and tcl8-doc-8.6.15-r0.riscv64
file /usr/share/man/mann/yield.ntcl conflicts between attempted installs of tcl-doc-9.0.0-r0.riscv64 and tcl8-doc-8.6.15-r0.riscv64
file /usr/share/man/mann/yieldto.ntcl conflicts between attempted installs of tcl-doc-9.0.0-r0.riscv64 and tcl8-doc-8.6.15-r0.riscv64
file /usr/share/man/mann/zlib.ntcl conflicts between attempted installs of tcl-doc-9.0.0-r0.riscv64 and tcl8-doc-8.6.15-r0.riscv64
[...]
Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Dan McGregor [Tue, 26 Nov 2024 18:06:42 +0000 (12:06 -0600)]
python3: package tkinter's shared objects
Seems every time python is updated this gets removed. That's likely
because the tk option can't be enabled in python3-native without
creating a dependency loop.
Instead of manually editing the manifest every time, explicitly
add it in the recipe as suggested by Alexander Kanavin.
Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Dan McGregor [Tue, 26 Nov 2024 18:06:41 +0000 (12:06 -0600)]
python3: do not overwrite FILES contents
Instead of relying completely on the python packaging manifest,
allow some default values. This is useful for the edge cases that
can't be found automatically by the create_manifest step.
In particular, tkinter can't be built for python3-native, so its
consistently missed in manifest updates.
Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 27 Nov 2024 14:50:00 +0000 (14:50 +0000)]
python3-flit-core: upgrade 3.9.0 -> 3.10.1
Summary of changes:
- The ``--python`` option can now take the path of a virtualenv folder, as an
alternative to a Python executable.
- The metadata in packages now has the names of optional dependency groups
("extras") normalised, complying with version 2.3 of the metadata standard.
- Fix potential substitution of environment variables into passwords read from
``.pypirc`` files.
- A warning is now shown when building packages which specify the old
``flit.buildapi`` backend, which should be replaced by
``flit_core.buildapi``. It's a good idea to always set a maximum
version for the build requirement, to protect against changes in
future major versions of Flit.
- Avoid using the deprecated ``datetime.utcfromtimestamp()``.
- The tests for ``flit_core`` are no longer part of the installed package,
reducing the size of the wheels.
Remove the -tests package, as the tests are no longer installed.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Trevor Gamblin [Wed, 27 Nov 2024 21:26:56 +0000 (16:26 -0500)]
python3-numpy: inherit pkgconfig
This was supposed to be added in oe-core commit fd656aaf7fbe
("python3-numpy: upgrade 1.26.4 -> 2.1.3"), but it somehow got missed in
the final version of that patch. Add it now to fix possible do_compile
errors on some systems, due to host paths sneaking into the meson build
options, e.g.:
| Build machine cpu family: x86_64
| Build machine cpu: x86_64
| Host machine cpu family: x86_64
| Host machine cpu: x86_64
| Target machine cpu family: x86_64
| Target machine cpu: x86_64
| Program python3 found: YES (/home/tgamblin/workspace/yocto/poky/build/tmp/work/core2-64-poky-linux-musl/python3-numpy/2.1.3/recipe-sysroot-native/usr/bin/nativepython3)
| Did not find pkg-config by name 'pkg-config'
| Found pkg-config: NO
| Run-time dependency python found: NO (tried pkgconfig, pkgconfig and sysconfig)
|
| ../numpy-2.1.3/meson.build:41:12: ERROR: Python dependency not found
|
| A full log can be found at /home/tgamblin/workspace/yocto/poky/build/tmp/work/core2-64-poky-linux-musl/python3-numpy/2.1.3/build/meson-logs/meson-log.txt
Richard Purdie [Wed, 13 Nov 2024 16:36:37 +0000 (16:36 +0000)]
insane: Split ERROR_QA into CHECKLAYER_REQUIRED_TESTS
We'd like to start requiring some set of checks are enabled for Yocto Project
Compatible Status and to pass yocto-check-layer. Start by splitting ERROR_QA
into two sets, the ones we think can be required and the ones we know have
challenges to implement (e.g. with prebuilt binaries).
To change the required list, the YP TSC would need to approve.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Wang Mingyu [Tue, 26 Nov 2024 01:29:58 +0000 (09:29 +0800)]
python3-websockets: upgrade 14.0 -> 14.1
Changelog:
============
- Supported max_queue=None in the asyncio and threading implementations for
consistency with the legacy implementation, even though this is never a good idea.
- Added close_code and close_reason attributes in the asyncio and threading
implementations for consistency with the legacy implementation.
- Once the connection is closed, messages previously received and buffered can
be read in the asyncio and threading implementations, just like in the legacy implementation.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Wang Mingyu [Tue, 26 Nov 2024 01:29:55 +0000 (09:29 +0800)]
ofono: upgrade 2.12 -> 2.13
Changelog:
==========
- Add support for handling QMI PIN and Lock methods.
- Add support for handling QMI WWAN interfaces.
- Add support for handling RMNet interfaces.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Wang Mingyu [Tue, 26 Nov 2024 01:29:51 +0000 (09:29 +0800)]
libsoup: upgrade 3.6.0 -> 3.6.1
Changelog:
============
* Fix `soup_uri_copy()` reading port as a long instead of an int
* Fix possible NULL deref in `soup_uri_decode_data_uri()`
* Fix possible overflow in `SoupContentSniffer`
* Fix assertion in `soup_uri_decode_data_uri()` on URLs with a path starting with `//`
* headers: Be more robust against invalid input when parsing params
* websocket: Fix possibility of being stuck in a read loop
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Wang Mingyu [Tue, 26 Nov 2024 01:29:49 +0000 (09:29 +0800)]
libadwaita: upgrade 1.6.1 -> 1.6.2
Changelog:
===========
- AdwDialog
- Fix accessible role critical when using window-backed dialogs
- AdwHeaderBar
- Fix a typo in docs
- AdwMessageDialog
- Fix a typo in adw_message_dialog_response() deprecation message
- AdwStyleManager
- Fix initial color scheme value on Windows
- Translation updates
- Dutch
- Serbian
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Wang Mingyu [Tue, 26 Nov 2024 01:29:47 +0000 (09:29 +0800)]
dnf: upgrade 4.21.1 -> 4.22.0
0005-Do-not-prepend-installroot-to-logdir.patch
refreshed for 4.22.0
Changelog:
=============
- doc: Naming of source and debug repos
- Fix /var/cache/dnf and /etc/yum packaging
- Allow --installroot on read-only bootc system
- spec: If DNF5 obsoletes DNF, do not build dnf and yum packages
- Print rpm unpack errors to the user (RhBug:2312906)
- package: remote_location() takes basedir into account
- spec: If DNF5 obsoletes DNF, do not build dnf-automatic
- Allow --downloadonly on read-only bootc system
- Add kernel-core to reboot_needed list
- Update need_reboot for dnf-automatic
- doc: Example send_error_messages in /etc/dnf/automatic.conf
- automatic: Check availability of config file
- Updated conf_ref to reflect change in fastestmirror behavior
- doc: Fix display issue of a code snippet
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Hiago De Franco [Tue, 26 Nov 2024 00:08:36 +0000 (21:08 -0300)]
weston: upgrade 13.0.3 -> 14.0.1
Key changes are:
- Patches have been removed since they were merged upstream and are no
longer needed on 14.0.1.
- Since commit a16598b038b2 ("backend-drm: make libdisplay-info
mandatory") from Weston, libdisplay-info is now a build dependency.
- Since commit f271dd34b617 ("compositor: delete cms-static and
cms-colord plugins") from Weston, "deprecated-color-management-colord"
has been removed, therefore remove the "colord" PACKAGECONFIG.
Signed-off-by: Hiago De Franco <hiago.franco@toradex.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Hiago De Franco [Tue, 26 Nov 2024 00:08:35 +0000 (21:08 -0300)]
libdisplay-info: add recipe
libdisplay-info is a EDID and DisplayID library. Goals:
- Provide a set of high-level, easy-to-use, opinionated functions as
well as low-level functions to access detailed information.
- Simplicity and correctness over performance and resource usage.
- Well-tested and fuzzed.
Since commit a16598b038b2 ("backend-drm: make libdisplay-info
mandatory") from Weston, this library is now mandatory in order to
build weston 14.0.1 and newer.
Signed-off-by: Hiago De Franco <hiago.franco@toradex.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Hiago De Franco [Tue, 26 Nov 2024 00:08:34 +0000 (21:08 -0300)]
hwdata: add recipe from meta-openembedded
Move hwdata recipe from meta-openembedded to openembedded-core.
libdisplay-info is now a requirement for Weston, since commit a16598b038b2 ("backend-drm: make libdisplay-info mandatory"). However,
in order to build libdisplay-info, hwdata is required. Since Weston is
built inside openembedded-core, move hwdata in order to be able to
upgrade Weston, so oe-core has no dependency on meta-oe.
hwdata contains various hardware identification and configuration data.
Signed-off-by: Hiago De Franco <hiago.franco@toradex.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
liburi-perl was added in 2010 to support lsb testing, never had
any other consumers in core, and since 2019 lsb testing is gone as well:
https://git.yoctoproject.org/poky/commit/?h=master-next&id=bac4bc9aa6a1f2fcf2ce9644925615185cc8e847
Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
intltool was dropped as a dependency in v236
See commit for reference:
https://github.com/systemd/systemd/pull/7313/commits/c81217920effddc93fb780cf8f9eb699d6fe1319
Signed-off-by: Guðni Már Gilbert <gudni.m.g@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
intltool was dropped as a dependency in v236
See commit for reference:
https://github.com/systemd/systemd/pull/7313/commits/c81217920effddc93fb780cf8f9eb699d6fe1319
Signed-off-by: Guðni Már Gilbert <gudni.m.g@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Peter Marko [Sun, 24 Nov 2024 12:37:58 +0000 (13:37 +0100)]
builder: set CVE_PRODUCT
Builder is a common word and there are many other builder components
which makes us to ignore CVEs for all of them.
There is already 1 ignored and currently 3 new ones.
Instead, set product to yocto to filter them.
Signed-off-by: Peter Marko <peter.marko@siemens.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Simone Weiß [Sat, 23 Nov 2024 17:43:22 +0000 (17:43 +0000)]
gnutls: Upgrade to 3.8.8
In 3.8.7 is that the linkage to compression libraries were changed to use a new
tool called dlwrap, this caused issues during the last upgrade but is resolved
now with the update to 3.8.8. Remove the workarounds for the update hence.
Changes in 3.8.8 include:
* libgnutls:
* Experimental support for X25519MLKEM768 and SecP256r1MLKEM768
key exchange in TLS 1.3
* All records included in an OCSP response are now checked in TLS
* Handling of malformed compress_certificate extension is now more standard
compliant
* More flexible library linking options for compression libraries, TPM, and
liboqs support. The configure options, --with-zstd, --with-brotli, --with-zlib,
--with-tpm2, and --with-liboqs now take 4 states:
yes/link/dlopen/no, to specify how the libraries are linked or
loaded.
Do not set a specific library linking option but use the default(yes).
Signed-off-by: Simone Weiß <simone.p.weiss@posteo.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joshua Watt [Fri, 22 Nov 2024 20:15:57 +0000 (13:15 -0700)]
lib/oe/sbom30: Prefix aliases with "http://spdx.org/spdxdocs/"
While the SPDX ID aliases are primarily intended for internal use by the
SPDX code, they are used in places where a valid SPDX ID is expected. In
order to make sure that they still conform to this, prefix them with
"http://spdx.org/spdxdocs/openembedded-alias"
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Fri, 22 Nov 2024 18:28:15 +0000 (18:28 +0000)]
gstreamer1.0: run ptests in parallel
Run the ptests in four parallel jobs (reduces runtime from ~80s to ~20s
on my machine), and also pass any arguments in the scripts to make it
easier to run individual tests.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Although support for SME was merged in v5.19, we've since uncovered a
number of issues with the implementation, including issues which might
corrupt the FPSIMD/SVE/SME state of arbitrary tasks. While there are
patches to address some of these issues, ongoing review has highlighted
additional functional problems, and more time is necessary to analyse
and fix these.
For now, mark SME as BROKEN in the hope that we can fix things properly
in the near future. As SME is an OPTIONAL part of ARMv9.2+, and there is
very little extant hardware, this should not adversely affect the vast
majority of users.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
]
Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Although support for SME was merged in v5.19, we've since uncovered a
number of issues with the implementation, including issues which might
corrupt the FPSIMD/SVE/SME state of arbitrary tasks. While there are
patches to address some of these issues, ongoing review has highlighted
additional functional problems, and more time is necessary to analyse
and fix these.
For now, mark SME as BROKEN in the hope that we can fix things properly
in the near future. As SME is an OPTIONAL part of ARMv9.2+, and there is
very little extant hardware, this should not adversely affect the vast
majority of users.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
]
Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Fri, 22 Nov 2024 11:07:13 +0000 (11:07 +0000)]
systemd: handle llvm-objcopy behaviour when reading .note.dlopen section
There are two behavioural differences between the objcopy in binutils
and llvm which resulted in build failures when building systemd with
clang:
1) If the section specified in --dump-section doesn't exist, binutils
set an exit code of 0 whereas llvm sets 1. This means we need to handle
the exit code so that we raise exceptions on unexpected failures, but
return an empty byte string if the segment isn't found.
2) binutils writes the section to the file name directly, whereas llvm
writes to a temporary file and renames. This means we can't read the
open fd directly, and instead need to re-open the file to read it.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joshua Watt [Fri, 22 Nov 2024 15:49:39 +0000 (08:49 -0700)]
spdx: Fix SPDX tasks not running when code changes
The SPDX code makes heavy use of python classes. While this works very
well, the bitbake dependency parser is unable to understand how to deal
with them, and thus changes to the class code do not cause rebuilds to
occur. To correct this, add the library files that include SPDX code as
file checksums for the SPDX tasks. If this method works well for SPDX,
we will look at implementing something similar in the bitbake dependency
parser that should allow correct checksums without having to explicitly
add them to each class.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Jörg Sommer [Fri, 22 Nov 2024 14:39:27 +0000 (15:39 +0100)]
python3-ctypes: depend on ldconfig only if distro-feature set
In the source code is only usage of ldconfig (for Linux) in
Lib/ctypes/util.py:301. This is wrapped in try-execpt which causes
_findSoname_ldconfig return Nothing. This is handled properly in
find_library and other methods to find the library a tried.
So, the code can handle the case of a missing /sbin/ldconfig (BTW:
hard-coded path). When DISTRO_FEATURES does not contain ldconfig some other
services (Systemd) to setup and maintain the ldconfig.cache are not
installed. Hence, this ldconfig pulled in by the ctype dependency is of no
use.
Therefore, do not set the dependency on ldconfig, if the distro-feature is
not set.
Signed-off-by: Jörg Sommer <joerg.sommer@navimatix.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Fri, 22 Nov 2024 14:21:58 +0000 (14:21 +0000)]
xserver-xorg: rewrite ABI dependency generation
This was motivated by remembering that both xserver-xorg and xorgxrdp
need to ignore the xorg-driver-abi test in do_package_qa because the
logic to generate the required dependencies is contained in
xorg-driver-common.inc, so can't be reused easily by the xserver (which
ships the modesetting driver) or xorgxrdp (which ships drivers and more).
Merge both the RPROVIDES (xserver) and RDEPENDS (driver) functions into a
single xserver-abi.inc to ensure that their logic remains in sync.
Generalise the names: instead of hardcoding 'input' and 'video' extract
the ABI names from the pkg-config file directly. This means 'input' is
now 'xinput' and 'video' is now 'videodrv', also 'ansic' and 'extension'
are new ABIs exposed.
Rewrite the RDEPENDS generation so that it is more flexible, and can be
used from inside the xserver-xorg recipe to generate RDEPENDS for the
modesetting driver. This means that recipe can remove the INSANE_SKIP.
There's an argument that this new .inc file could be a bbclass, I'm
undecided on this myself right now and this patch is essentially a
rationalisation of the existing code.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 22 Nov 2024 16:18:51 +0000 (16:18 +0000)]
resulttool: Improve repo layout for oeselftest results
Having all oe-selftest results on top of each other results in a large 640MB
json file which is hard to use. Split the results out per machine and test type.
This also stops the toolchain raw logs from overwriting each other meaning more
than one MACHINE is preserved.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 22 Nov 2024 16:17:19 +0000 (16:17 +0000)]
resulttool: Trim the precision of duration information
The duration values have pointless amounts of precision. Removing some of the
least significant digits reduces result size and makes the results easier to read.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 22 Nov 2024 16:15:46 +0000 (16:15 +0000)]
resulttool: Clean up repoducible build logs
We've improved the data stored for reproduicible builds. Teach resulttool how
to apply those cleanups when reprocessing data so we can reduce results file
sizes and make the data easier to process.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 22 Nov 2024 16:11:43 +0000 (16:11 +0000)]
resulttool: Add --logfile-archive option to store mode
Storing the log files inside the testresults git repo isn't scaling and isn't
really appropriate use of a git repository. Allow these to be optionally stored
in a separate filesystem location so the git repo can remain managable.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 21 Nov 2024 11:46:08 +0000 (11:46 +0000)]
resulttool: Use single space indentation in json output
Using 4 space indentation in resulted in hundreds of megabytes of extra file size
in general use. Reduce this to make filesizes more managable and reduce the processing
cost. Some level of indentation and spacing does make the files more readable and allows
use of git diff so we need to retain some of it.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Trevor Gamblin [Thu, 21 Nov 2024 15:38:59 +0000 (10:38 -0500)]
python3-numpy: upgrade 1.26.4 -> 2.1.3
- This release contains numerous API changes and bug fixes.
Changelog: https://github.com/numpy/numpy/releases
- Notably, the build backend has switched to mesonpy since 1.26.4.
- Patch 0001-numpy-core-Define-RISCV-32-support.patch had to be updated
to change the paths for the target files. It has now been merged
upstream, but isn't included in any release yet. PR for reference:
https://github.com/numpy/numpy/pull/17780
- Builds OK and seems to be compatible with current piglit and pandas
(meta-python), which was not the case for the previous upgrade attempt
to numpy 2.0.1.
- For numpy, add 'pkgconfig' to inherit to avoid errors like this at
do_compile:
| Found Pkg-config: NO
| Run-time dependency python found: YES 3.12
| Has header "Python.h" with dependency python: NO
|
| ../numpy-2.0.0/meson.build:44:2: ERROR: Problem encountered: Cannot compile `Python.h`. Perhaps you need to install python-dev|python-devel
|
| A full log can be found at /home/tgamblin/workspace/yocto/poky/build/tmp/work/core2-64-poky-linux/python3-numpy/2.0.0/build/meson-logs/meson-log.txt
- Mmodify the FILES:${PN}-staticdev line to fix a QA issue.
- This is now reproducible as of oe-core commit: 917df5ed022f
- Update tests to work with newly released pyproject-metadata 0.9.0.
Fix tests to work when not executed in a git work tree.
Daniele Nicolodi — 23-10-2024.
0.17.0
- Refuse to build wheels targeting the limited API and free-threaded
CPython: the free-threaded CPython build does not support the
limited API yet.
- Always use UTF-8 encoding for writing Meson native and cross
files. Always use UTF-8 encoding for reading read Meson
introspection data and pyproject.toml.
- Do not include uncommitted changes when creating source
distributions. Previously, uncommitted changes to files under
version control were included, but not untracked files. There was
no strong use case for this behavior and it was too surprising to
keep it.
- Make source distribution reproducible: use the modification time
of pyproject.toml for the generated files in the source
distribution archives.
- Disable the abi3 wheel tag for PyPy when building wheels targeting
the limited API: PyPy supports the limited API but not the stable
ABI.
- Raise ImportError when the package rebuild fails when importing an
editable install.
- Fix the wheel platform tag for GraalPy.
- Add .gitignore and .hgignore files to build directory if it is
empty.
- Allow install_subdir() of missing directories.
Christian Clauss, Daniele Nicolodi, Jonathan J. Helmus, Leo Singer, Loïc
Estève, Michael Simacek, Ralf Gommers, Simon McVittie — 19-10-2024.
Harish Sadineni [Wed, 20 Nov 2024 12:19:23 +0000 (04:19 -0800)]
linux: Modify kernel configuration to fix runqlat issue
In Qemu Running "/usr/share/bcc/tools/runqlat 1 5" gives following error:-
libbpf: kernel BTF is missing at '/sys/kernel/btf/vmlinux', was CONFIG_DEBUG_INFO_BTF enabled?
libbpf: failed to find valid kernel BTF
modprobe: FATAL: Module kheaders not found in directory /lib/modules/6.10.14-yocto-standard
Unable to find kernel headers. Try rebuilding kernel with CONFIG_IKHEADERS=m (module) or installing
the kernel development package for your running kernel version.
chdir(/lib/modules/6.10.14-yocto-standard/build): No such file or directory
Traceback (most recent call last):
File "/usr/share/bcc/tools/./runqlat", line 293, in <module>
b = BPF(text=bpf_text)
File "/usr/lib/python3.13/site-packages/bcc/__init__.py", line 479, in __init__
raise Exception("Failed to compile BPF module %s" % (src_file or "<text>"))
the following patch fixes the error.
BCC test summary for x86_64:
Before applying the patch:
TOTAL: 95
PASS: 29
FAIL: 66
After applying the patch:
TOTAL: 95
PASS: 79
FAIL: 16
Richard Purdie [Thu, 21 Nov 2024 14:55:59 +0000 (14:55 +0000)]
selftest/reproducible: Clean up pathnames
There are several problems with these paths. Firstly they contain full
system paths which depend upon where the test was run. These are pretty
pointless and just take up a lot of space making the results files large.
Secondly, they contain the same path twice. The reference and target path
will always be the same thing in two different locations.
Strip off the prefix and remove the duplication. This does change the output
data but that can't really be avoided. It does shrink the results data and makes
it more readable.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 21 Nov 2024 14:55:44 +0000 (14:55 +0000)]
selftest/reproducible: Drop rawlogs
The "rawlogs" data consists of a long string of results data which is
already in a structured data format. I can't see this is adding much
value in duplciating the data but it does create a huge string with a
lot of long problematic pathnames and inflates the results data size.
I suggest we drop this data as obsolete and not necessary.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
projects / thirdparty / openembedded / openembedded-core-contrib.git / log
