RMLTC0027b and RMLTC0027c

[chrdebru]

I can understand RMLTC0027a, where one wants to ensure that it generates URIs instead of IRIs, but, ideally, we should have:

• a test that would generate an IRI but not a URI that fails.

Could someone please explain the purpose of RMLTC0027b and c? The generated URIs/IRIs in output.nq are safe. Do you know if the output file is wrong?

[dachafra]

@chrdebru I was reviewing the 27b and 27c, and I agree they are wrong now.

My question is... does make sense to have a rml:UnsafeURI term type? Would the codification not be the same as with rml:UnsafeIRI? Taking the current example of the spec (http://data.example.com/author/{author} + "author": "Zoë Krüger"):

• IRI --> http://data.example.com/author/Zoë%20Krüger
• URI --> http://data.example.com/author/Zo%C3%AB%20Kr%C3%BCger
• UnsafeIRI --> <http://data.example.com/author/Zoë Krüger>
• UnsafeURI --> <http://data.example.com/author/Zoë Krüger>

Percentage-encoding

[dachafra]

fixed!