Apart from a few minor changes (adding color to debug outputs, making docker_run.sh and docker_build.sh work when called from outside the folder, ...), there are two bigger ones featured in this pull request:
The docker image now has its own version that may differ from the project version. The idea behind that is, that other files in the project change more often than the Dockerfile. With the old setup, every release also creates a new docker image - even if nothing did change there. It is now possible to create a new release for this project while still using the docker image created with a previous release.
The deploy_gardener.sh script (and the undeploy version too) has been made somewhat modular: it can be given an optional argument specifying from where to start. It then deploys that component and all components that come after it in the script, but not the ones before that step. This is useful for example if the deploy script crashed in the middle for whatever reason - now you can just fix the problem and continue the deployment from where it crashed.
Apart from a few minor changes (adding color to debug outputs, making docker_run.sh and docker_build.sh work when called from outside the folder, ...), there are two bigger ones featured in this pull request:
The docker image now has its own version that may differ from the project version. The idea behind that is, that other files in the project change more often than the Dockerfile. With the old setup, every release also creates a new docker image - even if nothing did change there. It is now possible to create a new release for this project while still using the docker image created with a previous release.
The deploy_gardener.sh script (and the undeploy version too) has been made somewhat modular: it can be given an optional argument specifying from where to start. It then deploys that component and all components that come after it in the script, but not the ones before that step. This is useful for example if the deploy script crashed in the middle for whatever reason - now you can just fix the problem and continue the deployment from where it crashed.