Hi Aaron,

Thank you for the report - it's just saved my a lot of debugging! I've come across the same thing elsewhere.

In 3.1.0

 tdb:GraphTDB    rdfs:subClassOf  ja:Model .

is not registered. Your workaround is good and safe when the bug fix is included.

JENA-1181

fixed in the next development build.

    Andy

On 15/05/16 13:04, Andy Seaborne wrote:
On 14/05/16 18:31, Aaron Coburn wrote:
Hello,

I just updated an existing Tomcat-based Fuseki deployment with the
latest 2.4.0 release. My previous setup used Fuseki 2.3.1 with the
OWLFBRuleReasoner enabled and a TDB-based dataset.

The exact configuration was quite similar to the example here:
https://github.com/apache/jena/blob/master/jena-fuseki2/examples/service-inference-2.ttl


 When starting up 2.4.0, however, the application would not deploy,
and I received this error:

org.apache.jena.assembler.exceptions.AmbiguousSpecificTypeException:
cannot find a most specific type for :tdbGraph, which has as
possibilities: ja:Model tdb:GraphTDB

I found that, by adding this statement to the configuration, the
application would load properly:

tdb:GraphTDB    rdfs:subClassOf  ja:Model .

So, while I now have this working properly, I wonder why tdb:GraphTDB
isn't already defined as an rdfs:subClassOf ja:Model (e.g. as
ja:MemoryModel is an rdfs:subClassOf ja:Model). That is, I am not
sure if this is a code error or a documentation issue.

Thanks, Aaron Coburn


Hi Aaron,

Can you send the exact configuration that behaves different in 2.3.1 and
2.4.0?  What about 2.3.0?  It shouldn't. It looks like exact details
matter here.

At 2.3.1 the way that "tdb:GraphTDB    rdfs:subClassOf  ja:Model ." gets
inserted automatically [1] is the same as 2.4.0 (well, "should be the
same"). It was different in 2.3.0.

     Andy



[1]

The

[] ja:loadClass "com.hp.hpl.jena.tdb.TDB" .
tdb:DatasetTDB  rdfs:subClassOf  ja:RDFDataset .
tdb:GraphTDB    rdfs:subClassOf  ja:Model .

triples are no necessart for TDB, jena-text or jena-spatial - they do
not need to be removed and should be harmless.


Reply via email to