Evaluate use of git-commit-id-plugin and make compatible with shallow clones or archives

[wetneb]

The use of the git-commit-id-plugin requires that Maven builds are triggered from a full clone of the repository. This makes the CI slightly more wasteful (since it has to fetch the entire repository instead of the commit being built). It also indicates that if someones downloads an archive of the source code from GitHub and tries to build it locally, it will probably also fail (because of the lack of a .git subdirectory).
The git-commit-id-plugin can be configured to be skipped if the information is not available, which is probably the right thing to do here.

[wetneb]

When this is done, the CI configuration should be changed to make all the clones shallow (remove the fetch-depth lines).