Hi Andy Andy Seaborne wrote: > On 11/10/11 21:57, Paolo Castagna wrote: >> Hi, >> while I was helping Jérôme on jena-users mailing list investigating his >> problem with LARQ SNAPSHOTs, I found another strange problem. >> >> Here is what I do: > .... > >> java -cp target/fuseki-0.2.1-SNAPSHOT-sys.jar tdb.tdbdump --loc=/tmp/tdb >> >> Exception in thread "main" java.lang.UnsupportedOperationException: >> subject cannot be null >> at com.hp.hpl.jena.graph.Triple.<init>(Triple.java:28) > > Something is out of date: > 1/ It's fixed (it's a different manifestation of JENA-9?)
I don't think it's a manifestation of JENA-9 https://issues.apache.org/jira/browse/JENA-9 Did you mean JENA-91? https://issues.apache.org/jira/browse/JENA-91 Probably. I saw the problem using org.apache.jena:jena-tdb:jar:0.9.0-incubating-SNAPSHOT. So, we should just publish a new 0.9.0-incubating-SNAPSHOT. I'll try that. > 2/ Even if not, that's not the error message: > > NodecSSE.decode has: > try { > Node n = t.asNode() ; > if ( n == null ) throw new TDBException("Not a node: "+str) ; > return n ; > } catch (RiotException ex) > { > throw new TDBException("Bad string for node: "+str) ; > } > > so a recent change means null is trapped early. (confused). > Maybe the builds are behind the codebase (I thought I'd updated > everything) - so timestamp checking might reveal something. We are in a transition phase (and we have two SNAPSHOTs for each module). When we will have just one SNAPSHOT per module, things will be better. Moreover, we can (and I think we should) set up nightly builds in Jenkins to publish a new SNAPSHOT each night (it's extremely easy to do that using the Apache Maven snapshot repository). I might just go ahead and do that. I can create parallel -nightly job just for this: publishing SNAPSHOTs on the Apache Maven snapshot repository. Not the old style. This way, as soon as we fix a bug... it gets rolled out in a new SNAPSHOT automatically. > > (can't check now - using offline email) Thanks for your reply. Paolo > > Andy
