Testing scripts locally

Most *scripts will run locally, or in a local docker container. (Exceptions may include iscript and notarization_poller, which are targeted to run on mac hardware.)

To test the *scripts fully locally, you need the secrets. Dep and dev *script pools tend to have low-security secrets, so these might be ok to use locally. You may also be able to create your own account(s), e.g. a Google Play account for pushapk testing, or a throwaway gpg key for signing testing.

For a given *script, these would be the steps to test on your laptop or docker container:

Create the script_config.yaml

Example script_config.yaml files can be found in each *script’s docker.d/worker.yml file. These are YAML with json-e config so they can support multiple configurations in the same file. You’ll want to resolve the json-e portions and just have a simple yaml or json config file.

Your work_dir and artifact_dir should be absolute paths. You may want to use $PWD/work and $PWD/artifacts.

Download a docker image

You can download a docker image built in taskcluster via either ./mach taskgraph load-image or taskgraph load-image, depending on whether you’re working from mozilla-central or from a standalone taskgraph virtualenv. If you’re able to either populate the right secrets via env vars, or start the container interactively, you may be good.

It’s also possible to use, say, a python 3 image and create a virtualenv manually.

Install the virtualenv

Create the virtualenv, e.g. pyenv virtualenv NAME; pyenv activate NAME

Next, install the scriptworker-scripts dependencies.

  • If your target *script’s setup.py contains scriptworker_client, you want to run cd scriptworker_client && python setup.py develop && cd ..

  • If your target *script’s setup.py contains mozbuild, you want to run cd vendored/mozbuild && python setup.py develop && cd ../..

Then install your *script. cd SCRIPTNAME && python setup.py develop && cd ..

Download a task to test

Assuming the *script is already live, you can download an existing task and modify it before running. You can do this by:

  • Make sure you’ve installed scriptworker in your virtualenv

  • scriptworker will provide a create_test_workdir helper tool.

    create_test_workdir --help  # for help
    create_test_workdir [--path PATH] [--overwrite] TASK_ID
    

    This will create a ./work directory, populate ./work/task.json with the task definition of task TASK_ID, and download any task.payload.upstreamArtifacts in ./work/cot/UPSTREAM_TASK_ID/PATH/TO/ARTIFACT

  • Optionally edit the task.json to test what you want to test. Optionally modify the contents of ./work/cot/... to test what you want to test.

  • Make sure your script_config.yaml is pointing at the same path for work_dir.

  • run SCRIPTNAME script_config.yaml | tee log to run the script against that task while capturing log output in a file and the console.