- From: Andy Seaborne <andy.seaborne@epimorphics.com>
- Date: Thu, 01 Nov 2012 08:09:50 +0000
- To: Gregg Kellogg <gregg@greggkellogg.net>
- CC: RDF WG <public-rdf-wg@w3.org>
On 01/11/12 01:35, Gregg Kellogg wrote: > Andy, I think there's a problem with this particular test [1] Gregg, Agreed. I'll change it as you suggest. In fact, all the prefixes should be http://example/ -- a large scale cut&paste error. Andy > > # colon is a legal pname character > @prefix : <http://example> . > @prefix x: <http://example> . > :a:b:c x:d:e:f :::: . > > Colon is indeed a legal pname character, but the resulting URI is not: <http://examplea:b:c>. > > Of course, this my depend on your definition of a syntax test. For me, this means to parse the input without generating an error. As I attempt to construct an IRI, the library I use (Ruby Addressable::URI) flags this as an issue. In this particular case, I believe it is because the authority part of the URI contains an illegal port. From 3986 and authority is defined as follows: > > authority = [ userinfo "@" ] host [ ":" port ] > > So this means, that the authority is broken down as follows: > > host - examplea > port - b:c > > and port is defined as > > port = *DIGIT > > I think the test would be just as valid if prefix were set to <http://example/>: > > # colon is a legal pname character > @prefix : <http://example/> . > @prefix x: <http://example/> . > :a:b:c x:d:e:f :::: . > > This would result in the following: > > <http://example/a:b:c> <http://example/d:e:f> <http://example/:::> . > > Gregg Kellogg > gregg@greggkellogg.net > > [1] http://svn.apache.org/repos/asf/jena/Experimental/riot-reader/testing/RIOT/Lang/Turtle/syn-prefix-06.ttl >
Received on Thursday, 1 November 2012 08:10:20 UTC