On 23/10/12 19:43, Michael Brunnbauer wrote:

hi all

here is the next problem with java-rdfa-0.4.2, jena-2.7.3 and
jena-fuseki-0.2.5-20121019.052315-69. This time, it is the language tag @2-ru
in http://www.fizkult-ura.com/fitness/35. I tried to load the attached .nt
file generated with java-rdfa-0.4.2 and jena-2.7.3 from that url to Fuseki.

> POST http://ts.foaf-search.net:3030/crawl/update

Is that a SPARQL Update, what exactly is it? Is it a LOAD -- or INSERT DATA?

The system recovered (no data loaded).



A minimal example would be:

-------------
<http://example/s> <http://example/p>  ""@2-ru .
-------------
and command:

riot --validate D.nt


Language tags can't have digits at that location.

rfc5646 / bcp47 has the grammar from hell in it but the WG has decided not adopt it (it's as big as turtle itself!)

Turtle uses

'@' [a-zA-Z]+ ('-' [a-zA-Z0-9]+)*

The first part is ascii-alpha only.

I just saw that some of my tools use SysRIOT.wireIntoJena() but the Crawler
does not. Would that help here if I catch the Exceptions ?

Yes.

Various things try to wire in RIOT but if you use only core Jena it might be (currently) possible to bypass the initialization. We've been discussing making jena-core pull in RIOT always via reflection, so we can do it now in advance of any code organsiation.



On a side note: I got problems with a tel: URI ending with space today from
Henry Story's homepage (already corrected by Henry):

  <a rel="foaf:phone" href="tel:+15106981206 ">

Yes - spaces are illegal in URIs for all URI schemes always.


java-rdfa-0.4.2 produced that URI with a trailing space in the graph and the
irichecker then refused it. Seems to be correct behaviour but Henry suggested
that the parser maybe should strip the space. Damian ?

Here is the Fuseki log for the first problem:

19:59:06 INFO  Fuseki               :: [190665] POST 
http://ts.foaf-search.net:3030/crawl/update
19:59:06 ERROR riot                 :: [line: 5, col: 204] Bad language tag
19:59:06 WARN  SPARQL_Update$HttpActionUpdate :: Transaction still active in 
endWriter - no commit or abort seen (forced abort)
19:59:06 WARN  Fuseki               :: [190665] RC = 500 : 
org.openjena.riot.RiotException: [line: 5, col: 204] Bad language tag
com.hp.hpl.jena.shared.JenaException: org.openjena.riot.RiotException: [line: 
5, col: 204] Bad language tag
        at 
org.openjena.riot.system.JenaReaderRIOT.readImpl(JenaReaderRIOT.java:150)

        Andy


Reply via email to