Pick up a patch make Vulkan CTS build on ARM hosts without NEON.
Reported-by: Ryan Eatmon <reatmon@ti.com>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Upgrade OpenGL ES CTS, picking up new tests and fixes for old ones.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Adding WITH_CAIRO=ON to fix this warning:
CMake Warning at libfreerdp/CMakeLists.txt:193 (message):
-DWITH_SWSCALE=OFF and -DWITH_CAIRO=OFF, compiling without image scaling
support!
Signed-off-by: Marc Ferland <marc.ferland@sonatest.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The WITH_ADD option does not exist. What was probably meant is
WITH_AAD (Azure AD auth).
Signed-off-by: Marc Ferland <marc.ferland@sonatest.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Upgrade to release 3.6.3:
- Make RE PCRE compatible
- Only execute Python interpreters
- fish: set variable scope to local to avoid clobbering global or
universal variables
- Documentation and help improvements
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Upgrade to release 3.2.0:
- Expose can_colorize as public API
- Add support for Python 3.15
- Drop support for Python 3.9
- Replace action-pre-commit-uv with prek-action
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The run-ptest script determines success by grepping for
"0 failed" in the test output. This could incorrectly
report success for cases like "10 failed" or "100 failed".
Update the script to rely on the test binary's exit status instead,
while still capturing and printing full test output for logging.
This makes the ptest behavior more robust and consistent
Signed-off-by: Nikhil R <nikhil.r@bmwtechworks.in>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
- make request and reconfigure behavior for lines requested by names
consistent with that of lines requested by offsets
- fix parsing non-tuple keys when requesting a line
- disable UP045 linter check
- use quoted types when casting
- generate wheels for CPython 3.14
- issue a warning when reconfiguring a line not in the original request
- update dependencies for wheel generation
- update licensing information
- coding style fixes
- minor performance improvements
License-Update: Follow the PEP 639 specification and remove the
deprecated license classifier in pyproject.toml. Make the license key
a top-level string. No changes in the license type.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Here is the setting in bpftool recipe:
COMPATIBLE_HOST = "(x86_64|aarch64|riscv64).*-linux"
bpftool only support 64 bit arch, so add it into NON_MULTILIB_RECIPES
to avoid expanding lib32- variant
Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
install step is not consistent especially when usrmerge is not
enabled ( which is default ), based upon what it finds on build host
it decides to install the xfsdump and xfsrestore locations to be either
/sbin or /usr/sbin, therefore it should be made consistent
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Add missing kernel module dependencies, fix a script that is trying to
run python instead of python3, and add a patch that fixes a test (submitted
upstream).
There are 2 notes:
1. For the tests to succeed the kernel must be compiled with CONFIG_NFT_TPROXY
2. There is 1 testcase that seems to be flaky: 0002-persist testcase randomly
fails. When it is comparing the test output with the expected output, sometimes
there is a comment present, and sometimes there isn't. Couldn't get to the
bottom of this.
The second point is the reason why the recipe remains on the
PTESTS_PROBLEMS_META_NETWORKING list.
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
1. Install missing test files
2. Enable network access by setting a nameserver in resolv.conf
because some tests require this
3. One test tries to ping Google through IPv6 - this test requires
ipv6 DISTRO_FEATURE. In case ipv6 is not enabled, skip this test
from the run-ptest script.
NB: the last mentioned test to work also requires IPv6 support from
your ISP. My ISP is IPv4-only, and I couldn't try it out. The above
comes from my reading of the testcode.
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The tests execute and pass successfully - I managed to run them
successfully, without a change on x86-64 and aarch64.
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The application should not be used in its current state:
it is not compatible with the current version of libxml[1].
Due to this, skip it for now - if it will be maintained again,
it can be re-enabled.
[1]: https://github.com/shlomif/perl-XML-LibXML/issues/84
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The tests are using only Python's unittest module, not pytest -
so adapt the run-ptest script to invoke unittest instead pytest.
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Add missing dependencies.
Also, fixing the tests have surfaced an actual bug: the module
expects unversioned perl library to be present on the system
(or at least present in Perl's $Config{libperl}), however the
OE Perl build has a versioned library, which causes final linking
to fail.
A patch to correct this is part of this change, and it has been
submitted upstream also.
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Also fix ptests.
Changelog:
1.26:
-Avoid test failure on RedHat by requiring DSA-SHA1 opt-in.
-Fix: Rocky linux 9: libcrypto error
1.25:
-Suppress AUTOLOAD for Net::DNS::SEC::Private::_index.
-Abandon support for OpenSSL pre-1.1.1 API.
-Fix: Tests exhibiting signatures with SHA-1 fail since Fedora 41
-Fix: Net::DNS minimum version dependency
1.24:
-Eliminate deprecated RSA CRT parameter names.
-Add support for SM3 digest.
1.23:
-Code refactoring of XS component.
1.22:
-Enable Ed25519 sign/verify for BoringSSL & LibreSSL.
-Use EC curve names instead of literal NIDs.
-Circumvent sign/verify test failures on EBCDIC platforms.
-Fix: openssl-SNAP-20230831 breaks RSA sign/verify
1.21:
-Add new t/TestToolkit.pm
-Rework pre-installation test scripts.
- Fix: libressl-3.7.1 breaks DSA verify
1.20:
-Improve Net::DNS::SEC::Keyset tests and error reporting.
-Avoid test failures if/when DSA|MD5|SHA1 become unsupported.
1.19:
-Discontinue support for obsolete ECC-GOST.
-Add LICENSE file to comply with Fedora/RedHat announcement
and WARNING of restrictions on use of strong cryptography.
1.18:
-Eliminate bareword filehandle usage.
-Eliminate indirect object syntax.
-Eliminate grep/map <expression>.
1.17:
-Recognise BIND private key accessed via symbolic link.
1.16:
-Improve testing of verify() functions.
-Rework code in Digest.pm
-SEC.xs code reduction.
1.15:
-Provide access to OpenSSL message digest implementations.
1.14:
-Improve exception capture in test scripts.
-Support more efficient algorithm mapping in Net::DNS.
1.13:
-Tweaks to resolve compilation errors with BoringSSL.
1.12:
-Avoid use of EC_POINT_set_affine_coordinates_GFp which is
deprecated in OpenSSL 3.0.0
-Reduce level of support for OpenSSL non-LTS releases.
1.11:
-Explain why compilation aborted in Net::DNS::SEC::DSA et al.
-Fix Makefile.PL to suppress parallel test execution.
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Also fix ptests.
Changelog:
-Suppress autovivified undefined $rr->{class} and $rr->{ttl}.
-Rework test scripts for SVCB and DELEG.
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Also fix ptests.
Added two backported patches that makes patches compatible with the latest openssl.
Changelog: https://metacpan.org/release/CHRISN/Net-SSLeay-1.94/changes
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
There are no reasons to build the demos and examples for the dynamic
lvgl library.
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Generating a configuration fragment for a feature can now be done using:
bitbake lvgl -c menuconfig && bitbake lvgl -c diffconfig
While this uses the standard cml1.bbclass, it expects that "make" is
used to run menuconfig etc. This is not true for lvgl, so unfortunately
the whole do_menuconfig() function had to be copied to be able to remove
the use of "make".
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This switches the recipe from manipulating the lv_conf.h file directly,
to instead using Kconfig to configure lvgl-demo-fb.
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This switches the recipe from manipulating the lv_conf.h file directly,
to instead using Kconfig to configure lvgl.
* Drop configuration of LV_TICK_CUSTOM* as they were removed in 9.0.0.
* Drop configuration options that match their defaults.
Co-authored-by: Adam x Nilsson <adam.x.nilsson@axis.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* Add +git to PV since there is no v9.4.0 tag for the
lv_port_linux_frame_buffer repository.
* Do not run the default do_install() as it will install all of lvgl.
* Remove the hack in do_configure:prepend() for SDL2_image as cmake
(with a little help from pkg-config) now finds SDL2_image correctly.
License-Update: Copyright year updated to 2025
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Drop manual installation of lv_conf.h. It has not been needed since at
least 9.2.2.
License-Update: Copyright year updated to 2025
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The descrec test fails currently, due to the different characterwidth
postgres uses depending on the current locale.
If the locale is C or posix, then each character is 1 byte. Otherwise,
if it's some UTF8 locale, then it is 4 bytes.
The test case expects 4 byte width by default, making it fail. This backported patch
allows for both 4 and 1 byte characters to be tested successfully.
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
