These are automated bots and testing that works on the Cockpit project. This includes updating operating system images, bringing in changes from other projects, releasing Cockpit and more.
The bots work in containers that are built in the cockpituous
repository. New dependencies should be added there in the tests/Dockerfile
file in that repository.
- The containers in the
cockpitous
repository invoke the.tasks
file at root of this repository. - The
.tasks
file prints out a list of possible tasks on standard out. - The printed tasks are sorted in alphabetical reverse order, and one of the first items in the list is executed.
Most bots are python scripts. Shared code is in the tasks/ directory.
Many bots file or work with issues in GitHub repository. We can use issues to tell bots what to do. Often certan bots will just file issues for tasks that are outstanding. And in many cases other bots will then perform those tasks.
These bots are listed in the ./issue-scan
file. They are written using the
tasks/__init__.py
code, and you can see example-task
for an
example of one.
The bot output is posted using the cockpitous sink code. See that link for how it works.
Development of the bots happens on GitHub at https://github.com/cockpit-project/bots/
There are static code and syntax checks which you should run often:
$ test/run
You will need to install the pyflakes and pycodestyle packages for python3 in order to run this script.
It is highly recommended to set this up as a git pre-push hook, to avoid pushing PRs that will fail on trivial errors:
$ ln -s ../../test/run .git/hooks/pre-push