W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > October to December 2012

Re: GSP progress

From: Chime Ogbuji <chimezie@gmail.com>
Date: Sun, 4 Nov 2012 19:01:49 -0500
To: Andy Seaborne <andy.seaborne@epimorphics.com>
Cc: SPARQL Working Group <public-rdf-dawg@w3.org>
Message-ID: <BF72C29820074FCB9589B60189DFFCFC@gmail.com>
See my response inline below.

On Sunday, November 4, 2012 at 4:08 PM, Andy Seaborne wrote:

> Chime,
> 
> Current results below, in EARL. Not quite finished but now very close.
> 
> POST of triples to the dataset does create graphs as per spec.
> 
> 1/ A couple of status code differences.
> 
> 2/ Multipart form upload isn't working in Fuseki
> 
> 3/ One mystery failure
> 
> "GET of POST - after noop" ==> unexpected returned RDF graph
> 
> 4/ Multipart POST is broken in Fuseki.
> 
> 
> Details:
> 
> ** 201/204
> One is "gsp:" -- maybe "PUT - replace empty graph" (which as no link 
> on the HTML page either_

I have added this one (and its followup) to the test document with the proper anchor:

http://www.w3.org/2009/sparql/docs/tests/data-sparql11/http-rdf-update/#put__replace_empty_graph
> This setup tracks empty graph so 204 is legal - the graph exists at that 
> point and no entity is returned.

The validator has been changed to accept 204 as well (per the spec).  204 has been added to accepted response codes 
in other cases below (as they are appropriate use of 204).  
> ** 200/204
> gsp:post__existing_graph
> 
> The POST returns no content so I believe 204 is legal.
Changed. 
> ** gsp:post__multipart_formdata
> code missing in Fuseki

Ok.  
> ** gsp:get_of_post__after_noop
> I don't understand this one

> This looks like a failure to detect the MIME type as text/turtle as 
> yesterday when you relaxed about test to look for contains 
> "text/turtle". But I have tried several formats and can't get it to 
> work. The response looks OK to me when I use curl.

> Can you give me a hint as to why it fails?

For that test, the response from the validator is "unexpected returned RDF graph". That error
is only returned if a) the content-type is not provided or it contains 'text/turtle' AND
b) the graph returned by the GET is not isomorphic to what is expected:

@prefix foaf: <http://xmlns.com/foaf/0.1/> . 
@prefix v: <http://www.w3.org/2006/vcard/ns#> . 

[] a foaf:Person; 
   foaf:businessCard [ a v:VCard; v:given-name "Alice" ] . 


So, it looks like it is detecting the MIME type properly but is not agreeing with the graph it is getting back. The last failure of this 
kind (from the validator logs, in order the tests were run), shows that the server returned the following to the validator:

body: empty
status: 200
content-length: 0

content-location: http://people.apache.org:3030/ds//1


content-type: text/turtle;charset=utf-8 
> The HTML formatting is wrong it says "/ds//1" but the server received 
> /ds/1 which is correct.

Hmm.  The output above is printed directly from what the http library is receiving from the GSP implementation. 
> Notes:
> "2012-11-04"^^xsd:dateTime
> that's an xsd:date, not a dateTime.

I have fixed this 
> There are 3
> earl:test gsp: .
> i.e. incomplete prefixed name.
> 
> and 3 tests with no href in the HTML form output.
I have updated the tests that were missing refs in the HTML and proper ids in the resulting EARL  
 
-- 
Chime Ogbuji
Sent with Sparrow (http://www.sparrowmailapp.com)
Received on Monday, 5 November 2012 00:02:20 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:49 GMT