Currently, Docker images are immediately push to Docker Hub and then tests are executed using these tagged versions.
If the image is broken, the "good" image is already overwritten. And the broken image is immediately visible to other pipelines and repositories.
OTOH, the image needs to be uploaded to a repository, otherwise the image can't be used in the pipeline as a container image.
Possible Solutions:
- The image could be tagged using the pipeline ID until finally tested. Then it get's renamed and pushed again.
- ...
Currently, Docker images are immediately push to Docker Hub and then tests are executed using these tagged versions.
If the image is broken, the "good" image is already overwritten. And the broken image is immediately visible to other pipelines and repositories.
OTOH, the image needs to be uploaded to a repository, otherwise the image can't be used in the pipeline as a container image.
Possible Solutions: