This introduces a new application option `--port`/`-p` to switch the
listening port from 3000 (default) to another port.
Docker image can be configured using the corresponding environment
variable `GITEA_SQ_BOT_PORT`.
Helm Chart allows setting `.Values.app.listeningPort`
Resolves: #25
Signed-off-by: Steven Kriegler <sk.bunsenbrenner@gmail.com>
Instead of re-inventing the wheel regarding configuration location
handling and validation, this introduces a new command flag `--config`
allowing for full flexibility of configuration filename and location.
This flag can also be defined via environment variable which allows
an easy way of starting the bot from command line, inside a Docker
container or using the Helm Chart.
It makes the custom environment lookup unnecessary and reduces some
complexity during startup and for writing tests.
Resolves: #10
Signed-off-by: Steven Kriegler <sk.bunsenbrenner@gmail.com>
The current code base regarding API entrypoint is not testable as it
directly connects to Gitea when creating the API endpoints. This
prevented my from writing tests in the past for that part.
As the SonarQube quality gate broke due to changes in the API entrypoint
logic, tests are now required to satisfy the quality gate.
Therefore, the instantiation of the API handlers is now decoupled from
building the bot API endpoints and follows the same interface wrapper
strategy as used for the Gitea API client. This makes it testable.
Now, tests are written for the most parts of the API entrypoint. I've
also noticed that there was much overhead within the tests for a
non-implemented function `fetchDetails`. So I dropped that function for
now.
Signed-off-by: Steven Kriegler <sk.bunsenbrenner@gmail.com>
Co-authored-by: justusbunsi <sk.bunsenbrenner@gmail.com>
Reviewed-on: https://codeberg.org/justusbunsi/gitea-sonarqube-bot/pulls/22
