Some datatype() and lang() tests

This thread on what datatype() returns died out before consensus was agreed:

http://lists.w3.org/Archives/Public/public-rdf-dawg/2005OctDec/thread#msg304
as witnessed by
http://lists.w3.org/Archives/Public/public-rdf-dawg/2005OctDec/0373

so I added some tests to make the discussion more concrete.

http://www.w3.org/2001/sw/DataAccess/tests/data/ExprBuiltins/


==== Added:

q-datatype-2.rq        Looks for things that have a datatype (not an error)
q-datatype-3.rq        Looks for things with datatype xsd:string
q-lang-1.rq            Looks for things with a langtag of any kind
q-lang-2.rq            Looks for things with a langtag of ''



I did some clearing up as well:

==== Fixed:

== data-built-1.rq
== result-datatype-1.rq

In the data, there was a literal of  1.0
Changed to 1.0e0
   (1.0 is or is becoming now a decimal in Turtle, not an XSD double)
Changed results to match.

== q-langMatches-3.rq
== q-langMatches-4.rq

lang("xyz", "*") does not match by SPARQL nowadays but these tests had it 
returning true.  rq23/24 says:

http://www.w3.org/2001/sw/DataAccess/rq23/rq24.html#func-langMatches
"""
The idiom langMatches( lang( ?v ), "*" ) will not match literals without a 
language tag as lang( ?v ) will return an empty string
"""

 Andy

Received on Tuesday, 22 August 2006 12:24:33 UTC