Removing expected output from RMLTC0004b-* test-cases?

[dachafra]

All of them have

Error expected: Yes
Expected knowledge graph: None, because an error is expected that stops the generation.

[DylanVanAssche]

I think this is a bigger problem...
IMO we need proper defined errors in the spec as well that are returned.

For example:

• Mapping rules invalid: return -1
• Data error: return -2
• ...

Of course with proper semantics ;)

The idea would be that you can always expect that an engine either returns triples or exits with a defined error code.

[dachafra]

Could be for the future, but at this moment we have certain confusion with these test-cases as they provide the output file in Github but following the spec they shouldn't, so a way to solve it now could be removing these files.

[DylanVanAssche]

@dachafra agreed.
It should be the same the R2RML ones are the RML ones.