Full test coverage of message specifications

[treiher]

The test coverage for all message specifications should be 100 %. Missing test data for a message should lead to an error instead of a warning.

[senier]

@treiher Is this a duplicate of Componolit/RecordFlux#1215?

[treiher]

Not really. Both are similar, but while Componolit/RecordFlux#1215 is about adding some test data for all specifications (i.e. achieving a test coverage of > 0 %), this issue is about reaching 100 % test coverage for all specifications. While the former is achievable with limited effort, I'm not sure if we can accomplish the latter in the near term.

[senier]

I see, so Componolit/RecordFlux#1215 is about doing something and Componolit/RecordFlux#1217 is about doing everything. Then we should keep both to track the long-term effort.