toRDF tli12 test seems inconsistent with the specification

tli12 test checks if list elements expanded to IRIs with a bad base.

Input:

{
  "@context": {
    "@base": "http://invalid/<>/",
    "list": {
      "@id": "foo:bar",
      "@container": "@list",
      "@type": "@id"
    }
  },
  "list": ["test"]
}

Output:

_:b0 <foo:bar> _:b1 .
_:b1 <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://www.w3.org/1999/02/22-rdf-syntax-ns#nil> .

Note that http://invalid/<>/ is not a valid IRI.

The context processing specification states that:
5.7.3) Otherwise, if value is an IRI, the base IRI of result is set to value.
5.7.5) Otherwise, an invalid base IRI error has been detected and processing is aborted.

Hence, it seems to me that parsing should fail with an invalid base IRI error instead of returning the expected RDF triples.

Start with the tli12 test and the supplied @base value, then compare its behavior with context-processing steps 5.7.3 and 5.7.5 in the JSON-LD API specification. Determine whether the test or implementation matches the invalid-base-IRI requirement, then run the relevant toRDF tests and confirm the expected failure or RDF output.