Age | Commit message (Collapse) | Author |
|
|
|
|
|
This script generates sorting keys so that 20240201-4 sorts before 20240201-20,
but the code assumes that there is only one hyphen when it splits.
However, there is now a patchstatus-meta-oe directory, which causes the
script to throw an exception.
Before padding, use a regex to check that the key is of the format we expect,
that is two integers separated by a hyphen.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This report visualises the AB-INT bugs over time, to help find trends
such as bugs which are no longer occuring, or bugs which should be fixed
urgently.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Currently the autobuilders seem to run at 25% idle at times. We see a lof
of bouncing on IO load and it appears the systems could cope with higher
values.
Increase the make load average, cpu and IO pressure limits a bit to
see if build times improve.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Add meta-exein yocto-check-layer testing. Due to HOSTTOOLS issues, we
need to had a horrible hack. v2 compatibility doesn't test for HOSTTOOLS
changes although perhaps the next version might.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This "git send-email" example is present in most Yocto/OE repos and is
quite useful for a quick copy/paste.
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The fixed config does match what all other lists do:
[yocto-autobuilder-helper][PATCH vX] XXX
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Since [0], ibus should be reproducible. So, re-enable it in
reproducibility tests.
[0]: https://git.openembedded.org/meta-openembedded/commit/?id=0ffbe2a79a5d49b2719c934ca0e73822fa3cbc2e
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Remove crossguid from reproducibility test: it packages a file containing
a full TMP path:
WARNING: crossguid-0.2.2-r0 do_package_qa: QA Issue: File /usr/share/crossguid/cmake/crossguid-config.cmake in package crossguid contains reference to TMPDIR [buildpaths]
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This is useful when using large sets of layers to
verify that the right layers have been added in
the right order.
Signed-off-by: Ross Burton <ross.burton@arm.com>
|
|
Extract the logging of auto.conf to a new log_file_contents() function,
and instead of calling it before _every_ call of bitbake, only show it when
actually writing the auto.conf.
Signed-off-by: Ross Burton <ross.burton@arm.com>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
There was a report of missing go-cross:populate_sysroot here:
https://autobuilder.yoctoproject.org/typhoon/#/builders/83/builds/6504/steps/23/logs/stdio
The builds that are expected to create the object (or take it from cache) are:
https://autobuilder.yoctoproject.org/typhoon/#/builders/42/builds/8481/steps/12/logs/stdio
https://autobuilder.yoctoproject.org/typhoon/#/builders/76/builds/8457/steps/25/logs/stdio
and neither mentioned that task anywhere, instead going directly to go-helloworld's rpm.
This means that go-cross can be neither taken from cache not written into it in
a-full, and we need '--runall build' to ensure it's not skipped over.
I also took the opportunity to match up the target image names with what the CDN
test actually checks.
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It is often helpful to know how many CVEs are open against a given recipe.
Add a summary table of this to the end of the CVE listing.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
For sysvinit with systemd, the systemd package will still be build.
This gives us a warning like below:
WARNING: systemd-1_255.1-r0 do_install: Using /home/root as root
user's home directory is not fully supported by systemd
Set ROOT_HOME to "/root" to avoid such warning.
Note that when using sysvinit as the init manager, /home/root is totally
valid, it's just that 'systemd' being in DISTRO_FEATURES causes systemd
to be built. So the only purpose of this patch's is to avoid warnings
on autobuilders.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The default INIT_MANAGER is 'sysvinit', to use systemd as the init
manager, we use INIT_MANAGER = 'systemd' because we can make use
of the settings in conf/distro/include/init-manager-systemd.inc.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Add a dry-run mode to be able to run send_qa_email locally but disabling
any output (no commit or tag pushed upstream, no mail sent). This eases
all release-related debugging. This dry-run mode is enabled by the
following changes:
- add a -d/--dry-run parameter to send_qa_email
- update test_results url to allow cloning test_results repository without
having its public key registered upstream
- skip test results storage
- do not erase test results temp dir
- skip email sending (but still generate it)
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Download links need to point to the latest point release.
Signed-off-by: Michael Halstead <mhalstead@linuxfoundation.org>
|
|
Update from mickledore to nanbield.
Signed-off-by: Michael Halstead <mhalstead@linuxfoundation.org>
|
|
generate_regression_reports is currently protect in a try/catch block to
prevent it from canceling QA email generation when encountering an issue,
but get_regression_base_and_target is not.
Make sure that get_regression_base_and_target can not prevent QA email from
being generated by adding it to the try/catch block protecting
send_qa_email. While doing so, make sure to preserve the exitcode variable
to make sure that the step is still marked as fail in autobuilder to make
sure the error does not go silent. However the variable is not needed as
global anymore since it is now used in a single function.
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
send_qa_email currently fails when dealing with release version starting a
new major, for example 5.0. This has been observed for example when trying
to generate 5.0_M1.rc1
This specific versioning makes previous tag computation method fall through
last branch which currently expects that the current release tag indeed
exists (5.0_M1), which is true when checking regression reports a
posteriori, but not in an autobuilder run (tag is added only when the
release has been "validated")
Fix tag computation for this case by getting previous release tag with git
ls-remote, instead of relying on git describe with a possibly non-existing
tag. While doing so, add a few tests about this specific case.
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
send_qa_email currently deals with tags in a "reproducible" way: despite
new versions being released on different branches, the computation of the
"reference" version for a specific input version always remain the same.
This behavior does not match perfectly real expectations: if at the some
point we get version 4.3.1 as a comparison reference for a regression
report, and 4.3.2 is released some time later, we want the next comparision
to be done against 4.3.2.
Start introducing this new behavior by allowing the tests to check returned
version against regex patterns instead of static strings, so we can for
example use wildcards on the "micro" version
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Use intermediate variables for test input/output, and print expected
output in subtest to get more info when a test fails
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The CDN tests fail since do_package is run without rpm's being generated.
With hash equivalence, this can lead to missing sstate artefacts from the
CDN.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
toaster build folder
Update run-toaster-test to include BUILDDIR and TOASTER_TEST_USE_SSTATE_MIRROR which will be use in the build test in toaster
Signed-off-by: Marlon Rodriguez Garcia <marlon.rodriguez-garcia@savoirfairelinux.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Strip quotes from directory variables as they caused path errors.
Add environment variable for the toaster artifact directory.
Migrate from tox and django tools to using pytest.
Install python module requirements from the script as this is no
longer handled by tox.
Signed-off-by: Alexander Lussier-Cullen <alexander.lussier-cullen@savoirfairelinux.com>
CC: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The older buildtools has a bug that zoneinfo is missing. Update to a newer
release version to fix this.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Publish what cleans up the public SSTATE to help with debugging.
Signed-off-by: Michael Halstead <mhalstead@linuxfoundation.org>
|
|
Pass the toaster test environment SSTATE_DIR and DL_DIR for faster
builds and TOASTER_DJANGO_TMPDIR to remove problematic temp files
from the root level '/tmp' directory.
Signed-off-by: Alexander Lussier-Cullen <alexander.lussier-cullen@savoirfairelinux.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Michael Halstead <mhalstead@linuxfoundation.org>
|
|
Sort versions as if we used semantic versioning.
Include updates for the 4.3 release as well.
Signed-off-by: Michael Halstead <mhalstead@linuxfoundation.org>
|
|
Some USERADD_UID/GID_TABLES were missing for lower layers which caused
some recipes to be not buildable because of skipped recipes (e.g. xfmpc,
python3-ldap)
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This is needed to fix a reproducibility error in netdata.
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Add a runner that installs the patchtest dependencies in a Python venv
and then starts patchtest's selftests.
Signed-off-by: Trevor Gamblin <tgamblin@baylibre.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Add a toaster test builder that runs the tox test suite using a new
run-toaster-tests script.
Signed-off-by: Alexander Lussier-Cullen <alexander.lussier-cullen@savoirfairelinux.com>
CC: richard.purdie@linuxfoundation.org
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Use the new OEQA_REPRODUCIBLE_EXCLUDED_PACKAGES to not trigger a test
failure on the known not-reproducible packages from meta-openembedded.
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
These produce the files at https://docs.yoctoproject.org/parsed-layers.json
and https://docs.yoctoproject.org/releases.json.
Signed-off-by: Michael Halstead <mhalstead@linuxfoundation.org>
|
|
Do another git-pull in the metrics repository before updating the
metrics, in case other metrics jobs running in parallel have updated the
repositories since they were cloned. There will always be possibility
of racing metrics jobs, but this should reduce the chance of it
happening.
An alternative would be to commit and then rebase before pushing, but I
fear that a git-merge could produce invalid JSON and we'd have to
manually fix up the repository. In my opinion, a wasted metrics run is
preferable to potentially corrupted repositories.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Do another git-pull in the metrics repository before updating the
metrics, in case other metrics jobs running in parallel have updated the
repositories since they were cloned. There will always be possibility
of racing metrics jobs, but this should reduce the chance of it
happening.
An alternative would be to commit and then rebase before pushing, but I
fear that a git-merge could produce invalid JSON and we'd have to
manually fix up the repository. In my opinion, a wasted metrics run is
preferable to potentially corrupted repositories.
[RP: Moved to after the bitbke invocation]
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
To do that we need a separate config file.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Add each a new builder "auh-meta-oe" and one step by meta-openembedded
layers.
Co-authored-by: David Pierret <david.pierret@smile.fr>
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
setup-auh and run-auh were doing what the AB config.json does:
* creating repo checkouts: Now use NEEDREPOS
* configuring bitbake env: Now use extravars
This refactoring is needed to prepare adding AUH meta-oe support.
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This writes the metrics to a new yocto-metrics-meta-oe repository, and
write the results to patchmetrics-meta-oe.
There's a great/ugly (delete to taste) hack to hide CVEs from oe-core,
so this only lists issues in meta-oe.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|