19.12.2020, von Christian Glahn
Having all the course material as well as our web-page on github provides one key benefit for digitizing our R&D as well as our educational processes: We can have a structured process around all these activities. git already gives us a basic distributed process. In github or gitlab we can then augment this process with additional steps easily.
These augmenting the processes means practically that we can include linting for all of our material. This does not have to get implemented as some kind of Bot, but is triggered by actions or pipelines. These steps are automatically run whenever something happens in a project or on the web-page. However, before we can use actions we need to understand our basic process.
- All projects have a
main-branch, which is the reference for the current development. - Nobody ever works on the
main-branch. Apart from me working out these steps. - Every human change to a project or the website requires an issue to handle. An issue can be that something is unclear, something is wrong, or something has just changed and needs updating. Automated processes may not create issues.
- Every change enters a pull request from a different branch or from the main branch of a different repository. Different repos are only allowed for people outside of our team.
- Pull requests will be tested for compliance, such as meta-data, linting or unit tests. If these tests pass, any pull request by team members will be automatically accepted and merged with the main branch.
Based on this basic process we aim at handling our projects.