Heptapod CI: (too) simple triggering of heptapod-tests

Instead of trying a programmatic conversion of branch names, (not clear whether that would apply to fill in the triggered pipeline branch), we're defining two jobs, using branch rules.

The drawback of this approach is that we're writing to the 'heptapod:testing' image over and over. With the current low traffic on Omnibus Heptapod, this should still be ok.

A followup could be to create Docker tags based on the Node id, and then trigger the functional tests by passing the Docker tag…