[redland-dev] Raptor Turtle RDFQ test

[Dave Beckett]

On Fri, 23 Apr 2004, Ian Davis wrote:
> Hi all,
> 
> I'm now running my parser against the raptor rdfq test
> (http://cvs.ilrt.org/cvsweb/redland/raptor/tests/turtle/rdfq-results.ttl).
> 
> It looks to me like this file is not valid Turtle since it ends with a
> semi-colon followed by a full stop:
> 
>     rs:solution
>         [ rdf:type rs:ResultSolution ;
>           rs:binding [ rs:variable "x" ;
>                       rs:value "anon2" ] ;
>           rs:binding [ rs:variable "y" ;
>                       rs:value _:a ]
>         ] ;
>     .

OK, well that test does seem to have a ; too much - it's not necessary - 
generates no new triples.
 
> Rapper (1.3.0) doesn't seem to be bothered by this but my parser
> barfs. My reading of the grammar precludes this syntax but maybe I'm
> wrong. Is it me or the test?

However, this is implementing the (at the time) Notation3 grammar
now at http://www.w3.org/DesignIssues/OldNotation3Grammar as the
old grammar. It allows an empty propertylist, i.e. nothing after
the ;

So... my quick reading is:
1) The example is valid N3 (and it was written as N3)
2) The Raptor code implements the N3 as defined above.
3) However, the Turtle grammar doesn't seem to allow it!

I better think about this some more about whether to
change the test, raptor and/or turtle.  Currently I'm
favouring fixing the turtle grammar.  Or checking what
the New N3 grammar says, see http://www.w3.org/DesignIssues/Notation3

Dave