Sample Definition and Criteria
Definition
A sample is a concise Zephyr application that provides an accessible overview of one or more features, subsystems, or modules. This includes kernel services, drivers, protocols, etc. Samples should be limited in scope and should focus only on demonstrating non-trivial or essential aspects of the subsystem(s) or module(s) being highlighted.
Samples are recommended when submitting new features; however, they are not mandatory.
Sample Criteria
1. Samples must not be used to test features or verify platforms
The primary purpose of a sample is to provide a reference to the user.
Samples must not use Zephyr’s testing framework.
Must not use
CONFIG_ZTEST
Must not use zasserts in samples.
If a sample can provide output that can be verified, then output should be evaluated against expected value to have some level of testing for the sample itself. Refer to Test Runner (Twister) for more details.
Although being able to run a sample successfully does verify the functionality is working as expected, this should not replace dedicated and comprehensive tests with full coverage to be submitted into the tests/ folder. In a sample, the only thing you test is buildability and, in some cases, if a sample is performing as expected, i.e. you are testing the sample, not the subsystem it builds on top.
2. Twister should be able to build every sample.
Every sample must have a YAML file. Reference: Test Runner (Twister).
For example:
tests: sample.kernel.cond_var: integration_platforms: - native_sim tags: - kernel - condition_variables harness: console harness_config: type: one_line regex: - ".*Waited and joined with 3 threads"Do not mark the test as build only. A sample should be able to build and run on the documented platforms. Use the
harness
option to skip the execution. Twister only attempts to flash and execute the build binary if the harness is set toztest
orconsole
.The default configuration for the sample must be in the
prj.conf
file and should be operational on all supported platforms. Do not rely on theextra_args
orextra_configs
options in the YAML file to build the sample.The tests should verify the default configuration of the sample and any optional features documented in the sample’s README file. Sample should never be used to test functionality of the subsystem or module the sample belongs to.
Any documented configuration options related to the sample and its operation can be verified using
extra_args
orextra_configs
options in the YAML file. Theprj.conf
file should have the base configuration options.Sample output can be validated by leveraging the
harness_config
regex option, wherever applicable.Use
platform_allow
to limit test to the platforms the sample was actually verified on. Those platforms should be listed in the sample’s README file.Make use of
integration_platforms
to limit the scope when there are timing or resource constraints. Ideally, one platform should be enough to verify the sample when changes are submitted to the Zephyr tree via a pull-request.Make the sample as generic as possible. Avoid making a sample platform specific unless it is for particular hardware.
3. A sample should provide sufficient coverage of a subsystem, feature, or module.
- DO’s:
Cover the most common and important use cases of the functionality.
Keep the code simple and easy to read. Example: samples/philosophers.
- DONT’s:
Samples must not test the negative or edge case behaviors.
Must not be unit tests.
4. Samples must be documented.
Samples must have a
README.rst
file in the samples folder. Example:samples/subsys/foo/README.rst
. clearly explaining the purpose of the sample, its hardware requirements, and the expected sample output, if applicable.Ensure that the
README.rst
file is accessible in the sample hierarchy starting atsamples/index.rst
.
- README Template:
Overview, if applicable.
Software/Hardware requirements
Building & Running instructions
Sample output, if applicable.
As a starting point, this sample is a good example to refer to samples/kernel/condition_variables/condvar.