Similar to "core-image-ptest: Switch to BBCLASSEXTEND parallel execution"
Refactor meta-oe-ptest-image.bb for an image per ptest in the form of
meta-oe-ptest-image-XXX using class extensions.
* Add meta-oe-ptest-all-image.bb depends on ALL meta-oe-ptest-image-XXX in meta-oe
* Add meta-oe-ptest-fast-image.bb depends on meta-oe ptests which take less than ~30s
This allows parallel execution of the tests.
The downside to this approach is the parsing time of an image generating
this number of class extensions is slow but making it easier to execute
and collect test data should outweigh that.
A useful advantage is that the dependencies of each ptest package are
checked/tested individually.
Add to local.conf:
IMAGE_CLASSES += "testimage"
All the meta-oe-ptest-image-XXX can be built with:
$ bitbake meta-oe-ptest-all-image
Subsequently, all the ptest images can be run with:
$ bitbake -c testimage meta-oe-ptest-all-image
You will likely want to limit the number of parallel image tests:
BB_NUMBER_THREADS = "4"
based on the number of TAP devices you have defined or the number of simultaneous
QEMU sessions your server can support.
Results can be summarized with:
$ resulttool report tmp/log/oeqa/testresult.json
Signed-off-by: Tim Orling <tim.orling@konsulko.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Similar to ptest-packagelists.inc in oe-core, this list enables
per image ptest (isolated ptests for a given recipe)
* Add recipes to PTEST_FAST_META_OE for ptests which take less than ~30s
* Add recipes to PTEST_SLOW_META_OE for ptests which take longer
* Add recipes to PTEST_PROBLEMS_META_OE for ptests which have failing tests
or other issues with ptest package
The initial list was taken from:
meta_oe_ptest_recipes=$(bitbake-layers show-recipes --recipes-only \
--layer meta-oe --inherits ptest --bare | tr '\n' ' ' | \
pcregrep -o1 '^NOTE:.+===(.+)$')
Signed-off-by: Tim Orling <tim.orling@konsulko.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
We should output the compiled executables into ${B} and install them
from it as well. Currently we put the build objects in the source
directory. While at it: don't assume we're already in the source
directory and instead prepend the source file with ${S}.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Fixes
checking for IPv6 compile-time support without -DINET6... ./configure: line 10004: syntax error near unexpected token `;;'
Signed-off-by: Khem Raj <raj.khem@gmail.com>
With version 3.22.11 it's not enough avoid linking to the
closed source ImageProcessor library, it must not be installed
either. This library is pre-stripped so package QA fails.
Add a new patch to fix installing HPLIP.conf into
/usr/share/ipp-usb/quirks. The upstream Makefile.am fails
to take DESTDIR into account and do_install fails because
of this. Also, instead of detecting the presence of the
target directory which is invalid for a source build,
create it and install the quirk file.
Signed-off-by: Zoltán Böszörményi <zboszor@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Add runtime dependencies for pcsc_scan and scriptor. Script
ATR_analysis is written in Perl and it is used to parse the smart
card ATR. This script is called by default by pcsc_scan.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The circular dependency was fixed in a different way in
meta-intel's openvino-inference-engine. Only some samples
in openvino use opencv and building the samples was made
optional.
Signed-off-by: Zoltán Böszörményi <zboszor@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
In postgresql sysview ptest are failing due to hidden build info in pg_config table.
The information is hidden due to existing patch 0001-config_info.c-not-expose-build-info.patch
So for passing the test we need to reduce the row count in the sysviews test.
Also for test results to be shown as pass we need to reduce the row count for
the expected count in the sysviews.out file.
Signed-off-by: Manoj Saun <manojsingh.saun@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Add a recipe for the simple program that allows to replicate android
reboot's behavior on linux (e.g. reboot bootloader, reboot recovery) as
well as test reboot mode drivers in the linux kernel.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
bwrap path is read from STAGING_BINDIR_NATIVE. This commit fixes reproducibility and runtime
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
ptest-runner is not available for Nodejs on target hence added run-ptest script
Signed-off-by: Archana Polampalli <archana.polampalli@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
==========
Fix buffer overread if soft-wrapping occurred and ASS_FEATURE_WHOLE_TEXT_LAYOUT was enabled via API or due to Encoding -1 (bug introduced in 0.16.0)
x86: detect GNU Hurd and configure NASM appropriately
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Support of new kernel features:
-----------------------------------
j1939: make use of new RX UAPI
isotp: rework FRAME_TXTIME handling and buffer size (Linux v5.18+)
isotp: add support for transmission without flow control (Linux v5.19+)
mcp251xfd-dump: add support for dumps created with Linux kernel >= v5.18
can-calc-bit-timing: import bit timing calculation algorithm from v5.19
cangen: add support for SO_TXTIME
includes: update to Linux 6.2 upstream (with CAN XL data structures)
Improvements and features:
--------------------------------
canfdtest: Add extended frame format / message length / FD / BRS support
can-calc-bit-timing: add support for calculation of CAN-FD bit timings
cangen: add option for absolute timeouts
cangen: new generation mode - partially randomized payload (-D)
slcan: increase BTR char limit to 8 for 32bit BTR
log asc converter: support len8_dlc for Classical CAN frames
candump: handle individual logfile name for stdout
cansniffer: add CAN FD support
cansniffer: make ASCII output switchable in CAN FD mode
canplayer: introduce option to limit the number of processed frames
Fixes:
-----------------------
candump: print number of dropped frames as unsigned values
candump: reserve enough space for ctrlmsg
candump: fix accidentally disabled traffic timeout feature
lib: fix seperator in snprintf_error_data
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
===========
- Core:
. Fixed incorrect check condition in type inference.
. Fix incorrect check in zend_internal_call_should_throw().
. Fixed overflow check in OnUpdateMemoryConsumption.
. Fixed bug GH-9916 (Entering shutdown sequence with a fiber suspended in a
Generator emits an unavoidable fatal error or crashes).
. Fixed bug GH-10437 (Segfault/assertion when using fibers in shutdown
function after bailout).
. Fixed SSA object type update for compound assignment opcodes.
. Fixed language scanner generation build.
. Fixed zend_update_static_property() calling zend_update_static_property_ex()
misleadingly with the wrong return type.
. Fix bug GH-10570 (Fixed unknown string hash on property fetch with integer
constant name).
. Fixed php_fopen_primary_script() call resulted on zend_destroy_file_handle()
freeing dangling pointers on the handle as it was uninitialized.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
They are not cross compile friendly as they poke into build host
directories to determine if system is upstart based and it can fail
build on such hosts e.g. ubuntu
If someone is still using upstart with OE then please fix it and send a
better fix upstream to re-enable it here.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Add PACKAGECONFIG[openvino]. It uses openvino-inference-engine-initial
as build dependency to avoid circular dependency.
Signed-off-by: Zoltán Böszörményi <zboszor@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The nodejs package contains a symbolic link `/usr/bin/corepack` to a
file from the nodejs-npm package.
```shell
build/tmp/work/core2-64-poky-linux/nodejs/18.14.2-r0/packages-split/nodejs/usr/bin$ ls -l
total 31932
lrwxrwxrwx 1 jan jan 45 Feb 21 05:08 corepack -> ../lib/node_modules/corepack/dist/corepack.js
-rwxr-xr-x 2 jan jan 32694304 Feb 21 05:08 node
build/tmp/work/core2-64-poky-linux/nodejs/18.14.2-r0/packages-split/nodejs-npm/usr/lib/node_modules/corepack/dist$ ls -l
total 2428
-rwxr-xr-x 2 jan jan 614216 Feb 21 05:08 corepack.js
```
Due to the link, the build system detects a dependency from package
`nodejs` to package `nodejs-npm`.
Since the `nodejs-npm` package depends on plenty more packages the link
should be moved to the `nodejs-npm` package.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
==========
* improve text-cursor (patch by Jan Engelhardt):
+ allow selecting CURSOR_BAR mode from command-line/Xresources.
+ draw cursor using filled rectangle instead of rectangle
outline to permit thicker underlines/bars.
+ scale up cursor relative to font size.
* improve readline modes (Fedora #2166860):
+ document readline modes
+ change the feature to configure by default
+ replace hard-coded SS3 for cursor movement with current mode
+ replace hard-coded erase/lnext characters with current values
* improve status-line (report by Thomas Wolff):
+ RIS turns off status-line
+ Right-margin (DECLRMM and DECSLRM) limits the length of text
written/updated in the status-line.
+ Most controls which affect the whole screen are ignored while
updating the status-line.
* modify configure check for tgetent to allow for some special cases
of ncurses configuration (report by Satadru Pramani).
* reduce timeout, improve warning message if resize is run on a
terminal which is not VT100-compatible.
* reduce compiler warnings in configure script.
0001-Add-configure-time-check-for-setsid.patch
refreshed for new version.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Switch the SRC_URI branch from master to stable-1.1
The 1.1.5 version was actually a bit ahead of that tag but
the ChangeLog comment for the two releases are:
Thu Feb 23 2023 Andreas Schneider <asn@cryptomilk.org>
* cmocka version 1.1.7
* Update ignore list for source tarball generation
Fri Feb 16 2023 Andreas Schneider <asn@cryptomilk.org>
* cmocka version 1.1.6
* Added new assert macros to compare 2 double given an epsilon
* Added meson build system
* Added header with version to TAP13 output
* Fixed issues with MSVC
* Fixed TAP output for skipped tests
* Fixed issue with fail_msg
* CMake generated configs for find_package(cmocka)
* Documentation improvements
CMake still seems to be the default build tool based on Install.md
so use that rather than the shiny Meson system.
One test, test_cmockery, was added so now there are 23 passing ptests.
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
ChangeLog:
https://github.com/CESNET/libyang/releases/tag/v2.1.55
* Generate cases list dynamically in run-ptest.
* Add a patch to fix ptest.
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Upgrade urgency: SECURITY, contains fixes to security issues.
Security Fixes:
* (CVE-2023-28425) Specially crafted MSETNX command can lead to assertion and denial-of-service
Bug Fixes
=========
* Large blocks of replica client output buffer may lead to psync loops and unnecessary memory usage (#11666)
* Fix CLIENT REPLY OFF|SKIP to not silence push notifications (#11875)
* Trim excessive memory usage in stream nodes when exceeding `stream-node-max-bytes` (#11885)
* Fix module RM_Call commands failing with OOM when maxmemory is changed to zero (#11319)
Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Under sysvinit when trying to restart redis-server using
service redis-server restart two calls are made to start-stop-daemon, first
with the --stop argument and then with --start argument consecutively.
Because the process doesn't immediately terminate when start-stop-daemon
--stop is called, the next call to start-stop-daemon --start finds the
process still running and does not attempt to start another one.
This leads to only a stop of the redis-server process when a restart is
requested. This behavior affects all redis versions using sysvinit only.
This can be fixed by using the --retry <timeout/schedule> argument with
start-stop-daemon --stop in order for the call to block until the process
terminates so that start-stop-daemon --start will attempt to start a new
process.
Unfortunately the --retry argument works only in the implementation of
start-stop-daemon provided by dpkg package and is ignored in the
implementation provided by busybox package.
A repeated check if the process is still running and another try with
another signal after a timeout will effectively simulate a stop with
--retry=TERM/5/KILL/5 schedule.
Signed-off-by: Stefan Ghinea <stefan.ghinea@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
We get a very cryptic error during do_package
ERROR: Didn't find service unit 'lircmd.service', specified in SYSTEMD_SERVICE:lirc.
The reason is on ubuntu build hosts configure pokes at build host and
decided its ubuntu and applies hardcoded value of /lib/systemd/system
for systemdunitdir instead of /usr/lib/systemd/system as passed from
recipe
Signed-off-by: Khem Raj <raj.khem@gmail.com>
