Robot Framework in a lightweight Alpine Linux based Docker container.

1. To take Robot Framework into use without (having sudo to do) installations on the host
2. To avoid virtualenvs, etc. to deal with conflicting dependencies of the test libraries
3. To be able to install Robot Framework and run the tests with one command

Download rfdocker and Dockerfile to where your tests/ directory is:

curl https://raw.githubusercontent.com/asyrjasalo/rfdocker/master/rfdocker -o rfdocker || \
wget https://raw.githubusercontent.com/asyrjasalo/rfdocker/master/rfdocker -O rfdocker && \
chmod +x rfdocker && \
curl https://raw.githubusercontent.com/asyrjasalo/rfdocker/master/Dockerfile -o Dockerfile || \
wget https://raw.githubusercontent.com/asyrjasalo/rfdocker/master/Dockerfile -O Dockerfile && \
echo "Done."

Then:

./rfdocker

The Robot Framework output files are put in the same directory under results/.

Any arguments are forwarded to robot inside the container, e.g. the output directory can be renamed with:

./rfdocker --outputdir results/$(date +%Y-%m-%dT%H:%M:%S) tests/

Put the external test libraries to requirements.txt and uncomment the lines in Dockerfile. The packages are automatically installed inside the container whenever rfdocker is ran.

If the external test libraries require OS-level dependencies, you may need to add installation of these to Dockerfile (before running pip).

On Alpine Linux, you can install packages with RUN apk add --update {{packagename}}. See Alpine Linux wiki page for more information, and docker/Dockerfile.base in this repository as an example.

You can pass variable BUILD_ARGS to rfdocker to customize docker build arguments, and variable BUILD_DIR to override the path of the directory where Dockerfile and requirements.txt are.

Similarly, you can pass variable RUN_ARGS to rfdocker to e.g. define additional volume mappings for CI, to persist results/ outside the Jenkins job's workspace, or to allow connecting to your host machine from the container (RUN_ARGS="--network=host").

Git branches 'master' and 'python2' are to build images for Robot Framework on Python 3 and 2 series, respectively. The both images are hosted in Docker Hub, robotframework/rfdocker. The version numbers correspond to the Robot Framework releases.

In addition to version numbered tags, tag 'latest' references to the latest Robot Framework final (no rcs) on Python 3 series - please use tag 'latest' carefully though, in general with Docker images, as it is possible that the upcoming Robot Framework versions could introduce some backwards incompatible changes in the future, without you noticing this as easily as when using explicit version numbers.

The images are built using rfdocker itself, with the help of scripts/. You may benefit from these scripts for distributing your own images as well.

To build a new image:

scripts/build

The Robot Framework version is read from file scripts/release_name, and Python version from file scripts/python_version.

To push the image to Docker registry (and create a git tag after release_name):

scripts/release https://your.private.registry.com:5000/rfdocker

For pushing to Docker Hub, you can use:

scripts/release {{organization}}/rfdocker