[
https://issues.apache.org/jira/browse/JENA-471?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13684130#comment-13684130
]
Andy Seaborne commented on JENA-471:
------------------------------------
The 2.10.0 algebra is correct. The Nesting (project) is harmless.
Fix added to trunk - the problem was that there are two uses of the visitor,
one for visible variables and one for all mentioned variables. The latter is
used by (graph) to check for use of ?g both in GRAPH and in the inner pattern,
where the code ha to be careful when converting to quads.
Would it be possible to contribute the additional tests you have to the
project? Then this sort of thing will be tested during development.
> Regression in Algebra.toQuadForm() for nested sub-query with GRAPH clause
> -------------------------------------------------------------------------
>
> Key: JENA-471
> URL: https://issues.apache.org/jira/browse/JENA-471
> Project: Apache Jena
> Issue Type: Bug
> Components: ARQ
> Affects Versions: Jena 2.10.1
> Reporter: Rob Vesse
>
> Our internal testing has flagged up a regression in behaviour between 2.10.0
> and 2.10.1 with regards to Algebra.toQuadForm() when applied to sub-queries
> nested within a GRAPH clause
> The query in question is as follows:
> {noformat}
> SELECT ?x
> WHERE
> { GRAPH ?g
> { { SELECT ?x
> WHERE
> { ?x ?p ?g }
> }
> }
> }
> {noformat}
> With 2.10.0 this produced the following algebra:
> {noformat}
> (project (?x)
> (project (?x)
> (quadpattern (quad ?g ?x ?p ?g))))
> {noformat}
> With 2.10.1 this now produces an exception:
> {noformat}
> com.hp.hpl.jena.sparql.ARQInternalErrorException
> at
> com.hp.hpl.jena.sparql.algebra.OpVars$OpVarsPattern.visit(OpVars.java:197)
> at com.hp.hpl.jena.sparql.algebra.op.OpProject.visit(OpProject.java:48)
> at
> com.hp.hpl.jena.sparql.algebra.OpWalker$WalkerVisitor.visit1(OpWalker.java:86)
> at
> com.hp.hpl.jena.sparql.algebra.OpVisitorByType.visitModifer(OpVisitorByType.java:42)
> at
> com.hp.hpl.jena.sparql.algebra.OpVisitorByType.visit(OpVisitorByType.java:154)
> at com.hp.hpl.jena.sparql.algebra.op.OpProject.visit(OpProject.java:48)
> at com.hp.hpl.jena.sparql.algebra.OpWalker.walk(OpWalker.java:43)
> at com.hp.hpl.jena.sparql.algebra.OpWalker.walk(OpWalker.java:33)
> at com.hp.hpl.jena.sparql.algebra.OpVars.mentionedVars(OpVars.java:70)
> at com.hp.hpl.jena.sparql.algebra.OpVars.mentionedVars(OpVars.java:62)
> at
> com.hp.hpl.jena.sparql.algebra.AlgebraQuad$Pusher.visit(AlgebraQuad.java:93)
> at com.hp.hpl.jena.sparql.algebra.op.OpGraph.visit(OpGraph.java:46)
> at
> com.hp.hpl.jena.sparql.algebra.OpWalker$WalkerVisitor.before(OpWalker.java:64)
> at
> com.hp.hpl.jena.sparql.algebra.OpWalker$WalkerVisitor.visit1(OpWalker.java:84)
> at
> com.hp.hpl.jena.sparql.algebra.OpVisitorByType.visit(OpVisitorByType.java:110)
> at com.hp.hpl.jena.sparql.algebra.op.OpGraph.visit(OpGraph.java:46)
> at
> com.hp.hpl.jena.sparql.algebra.OpWalker$WalkerVisitor.visit1(OpWalker.java:85)
> at
> com.hp.hpl.jena.sparql.algebra.OpVisitorByType.visitModifer(OpVisitorByType.java:42)
> at
> com.hp.hpl.jena.sparql.algebra.OpVisitorByType.visit(OpVisitorByType.java:154)
> at com.hp.hpl.jena.sparql.algebra.op.OpProject.visit(OpProject.java:48)
> at com.hp.hpl.jena.sparql.algebra.OpWalker.walk(OpWalker.java:43)
> at com.hp.hpl.jena.sparql.algebra.OpWalker.walk(OpWalker.java:38)
> at
> com.hp.hpl.jena.sparql.algebra.Transformer.applyTransformation(Transformer.java:156)
> at
> com.hp.hpl.jena.sparql.algebra.Transformer.transformation(Transformer.java:149)
> at
> com.hp.hpl.jena.sparql.algebra.Transformer.transformation(Transformer.java:138)
> at
> com.hp.hpl.jena.sparql.algebra.Transformer.transformation(Transformer.java:132)
> at
> com.hp.hpl.jena.sparql.algebra.Transformer.transform(Transformer.java:57)
> at
> com.hp.hpl.jena.sparql.algebra.Transformer.transformSkipService(Transformer.java:87)
> at
> com.hp.hpl.jena.sparql.algebra.AlgebraQuad.quadize(AlgebraQuad.java:56)
> at com.hp.hpl.jena.sparql.algebra.Algebra.toQuadForm(Algebra.java:89)
> {noformat}
> I assume this regression happens because of the change to AlgebraQuad to use
> OpVars.mentionedVars() in the 2.10.1 dev cycle
> While the 2.10.0 algebra may have actually been incorrect it didn't throw a
> stack trace and I assume it is possible to make a valid quad transform on
> this algebra.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
