- From: Gregg Kellogg <gregg@greggkellogg.net>
- Date: Mon, 12 Oct 2015 09:57:35 -0700
- To: Andy Seaborne <andy@apache.org>
- Cc: public-rdf-tests@w3.org
Gregg Kellogg gregg@greggkellogg.net > On Oct 12, 2015, at 1:02 AM, Andy Seaborne <andy@apache.org> wrote: > > On 11/10/15 19:41, Ruben Verborgh wrote: >> Hi Gregg, >> >>> * Add Turtle, Trig, RDF/XML and SPARQL tests to better validate relative IRI resolution >> >> In the pipeline (tests are ready, just need to put them in the right format). >> >> Another thing I remember once being mentioned on the mailing list: >> extra syntax validations, for instance, that no space is allowed between literals and their type/language. > > Which way round are you saying that there needs to be validation? To allow whitespace or forbid whitespace? > > For better of worse and without judgement: the grammar says: > > [128s] RDFLiteral ::= String (LANGTAG | '^^' iri)? > > and it is in the part where whitespace between tokens is allowed > and LANGTAG is a token so no whitespace internally. > > So it's a bit inconsistent: > > :legal :p "foo" @bar . > :illegal :p "foo"@ bar . > > :legal :p "foo" ^^ :datatype . Yes, having tests to look both where whitespace is allowed, and where it is forbidden would be useful. The EBNF grammar specifically allows whitespace between tokens, but not within a token, unless it is explicitly defined. Gregg > Andy > >> >> Best, >> >> Ruben >> > >
Received on Monday, 12 October 2015 16:58:04 UTC