ZEPHYR_SDK_INSTALL_DIR is a required environment variable, otherwise the
script fails.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
This commit updates the Bluetooth test workflow to fetch all branches
during checkout by specifying the fetch depth of 0, in order to allow
rebasing onto the latest main branch before running the workflow.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
The 'bluetooth-tests' workflow is currently written such that it
only triggers on the pull requests that make changes to the files under
specific paths, which do not include the workflow itself.
This allows the workflow changes to be merged without being tested in
the CI and may lead to CI breakages when such changes contain errors.
This commit adds the workflow itself to the list of trigger paths so
that any changes made to the 'bluetooth-tests.yaml' and
'bluetooth-tests-publish.yaml' workflows are tested in the CI before
merging.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
Like in most other actions, rebase the branch on to of the latest main
branch in order to ensure that we pick up any hotfixes.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
With this setting enabled, Git credentials are not kept after checkout.
Credentials are not necessary after the checkout step since we do not
do any further manual push/pull operations.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Split Bluetooth tests workflow into 2 steps:
- One that runs the actual tests and stored results
- A second one that fecthes and uploads tests results
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Make sure sources are up-to-date before installing any package. This is
causing failures on some workflows.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The -q flag supresses all wget output, even when there is a failure.
This makes it difficult to diagnose CI failures when the download fails.
The --no-verbose flag is a better choice: it is silent enough but it
always shows the outcome.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The documentation supports a special target named "html-fast" that skips
generation of all Kconfig pages. Instead, it creates a single dummy page
where a reference to all existing Kconfig options is placed. This means
that references are resolved, but content is not rendered. Since Kconfig
help is rendered as a literal, chances of breaking documentation
build due to Kconfig changes should be low. The change proposed in this
patch should speed up documentation build on pull requests while a
proper solution is found for the Kconfig docs.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Add the documentation build jobs to a concurrency group so that branch
force pushes will automatically cancel in progress jobs.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
When an error happens during the Sphinx build (e.g. due to a broken
reference), the process hangs on CI when run with both `-j auto`
(parallel build) and `-W` (warnings as errors) options. The root cause
of the issue is unknown, and does not seem to happen locally. Parallel
builds are experimental on Sphinx, so they have been disabled on CI for
now. Because CI runner is a single core machine, the build time should
remain equal or similar. The option is still left as default, so local
builds will continue to benefit from parallelization.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
We are reusing already checked out trees that might be in the middle of
a rebase or conflict. Remove files preventing us from doing another
rebase after the tree was reset.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Merge all results from all nodes and generate an HTML report and upload
it as an artifact for review of failures offline.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
If variable exit is not defined, the test is not properly executed.
Note that in case of an actual error the var would be defined and the
test executed properly.
Signed-off-by: Nicolas VINCENT <nicolas.vincent@vossloh.com>
Improve calculation of matrix and move calculations from workflow to the
testplan script. We now generate a file that can be parsed by the action
with the data needed to start twister with the right number of nodes.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
When the number of tests is smaller, but not too small, we still run on
10 builders, if the test is small enough however, we will determine the
number of builders automatically.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Pull in the CI docker image v0.21.0, which contains the Zephyr SDK
0.13.2 release, and use the Zephyr SDK 0.13.2 for building and testing
Zephyr in the CI.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
Merged 3 files used to generate the test plan per PR based on the
changes. 2 python scripts and a shell script are now all merged into 1
python script that generates the input file for twister based on a list
of changed files by the PR.
This remove lots of old and obsolete code and simplifies things a bit,
no need anymore for an intermediate script to call twister, we call it
directly in the workflow and use the new test_plan script to generate
the test plan.
This also reenables the recently disabled tag based filtering which had
a bug, bug is resolved in this new implementation.
On push events, we now run twister without the --integration option to
catch any issues in the main branch that were not caught in PRs. PRs
continue to run with --integration enabled. This event (push) is now run
on 15 builders due to the increased size.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Give documentation build up to 30 minutes to finish. This should improve
the user experience when Sphinx hangs due to broken references, a
problem that needs some investigation.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
This action replaces current buildkite workflow and uses github actions
to build and run tests in the zephyr tree using twister. The main
differences to current builtkite workflow:
- the action handles all 3 events: pull requests, push and schedule
- the action determines size of matrix (number of build hosts) based on
the change with a minimum of 1 builder. If more tests are built/run
due to changes to boards or tests/samples, the matrix size is
increased. This will avoid timeouts when running over capacity due to
board/test changes.
- We use ccache and store cache files on amazon S3 for more flexibility
- Results are collected per build host and merged in the final step and
failures are posted into github action check runs.
- It runs on more powerful instances that can handle more load.
Currently we have 10 build hosts per run (that can increase depending
on number of tests run) and can deliver results within 1 hour.
- the action can deal with non code changes and will not allocate more
than required to deal with changes to documentation and other files
that do not require running twister
The goal long-term is better integrate this workflow with other actions
and not run unncessarily if other workflows have failed, for example, if
commit message is bogus, we should stop at that check, to avoid wasting
resources given that the commit message will have to be fixed anyways
which would later trigger another run on the same code.
Currently there is 1 open issue with this action related to a github
workflow bug where the final results are not posted to the same workflow
and might appear under other workflows. Github is working on this bug.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Process issues are tracked and do not need to be automatically made
stale.
Suggested-by: Anas Nashif <anas.nashif@intel.com>
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
