- From: Philip Taylor <pjt47@cam.ac.uk>
- Date: Wed, 11 Nov 2009 09:25:21 +0000
- To: Manu Sporny <msporny@digitalbazaar.com>
- CC: RDFa mailing list <public-rdf-in-xhtml-tf@w3.org>
Manu Sporny wrote:
> We had said that TC 140 shouldn't generate the triple listed in the
> SPARQL, but I can't remember why now (and it wasn't minuted). TC140 has
> been published as a negative test, but Philip meant it to be a positive
> test. Why should TC 140 not generate the triple listed in the SPARQL?
If it's agreed that no triple should be generated for this test case,
then the SPARQL test is inadequate for finding implementation bugs. In
particular, librdfa generates the triple:
<http://rdfa.digitalbazaar.com/test-suite/test-cases/xhtml1/0140.xhtml>
<http://rdfa.digitalbazaar.com/test-suite/test-cases/xhtml1/0140.xhtml_>
"Test"
which looks like a clear error, yet the test case does not detect it
since it's only testing for absence of one specific triple. Some other
implementations have different errors (treating it like a blank node).
I would suggest changing the query to say:
# This is a negative test, no triples may be generated.
ASK WHERE {
?s ?p ?o .
}
--
Philip Taylor
pjt47@cam.ac.uk
Received on Wednesday, 11 November 2009 09:25:58 UTC