https://issues.apache.org/jira/browse/TINKERPOP-2095
This PR fixes a bug that somehow managed to stay unnoticed for years.
Prior this PR we got this:
```
gremlin>
g.V().hasLabel('person').as('p').out('created').group().by('name').by(select('p').values('age').sum())
java.lang.Long cannot be cast to
org.apache.tinkerpop.gremlin.process.traversal.traverser.util.TraverserSet
```
Now we get this:
```
gremlin>
g.V().hasLabel('person').as('p').out('created').group().by('name').by(select('p').values('age').sum())
==>[ripple:32,lop:96]
```
The fix was made for both - `GroupStep` and `GroupSideEffectStep`. I also had
to fix `ComputerVerificationStrategy`; it didn't detect that
`select('p').values('age')` is leaving the star-graph.
Without the `ComputerVerificationStrategy` fix:
```
gremlin>
g.V().hasLabel('person').as('p').out('created').group().by('name').by(select('p').values('age').sum())
==>[ripple:0,lop:0]
```
With the `ComputerVerificationStrategy` fix:
```
gremlin>
g.V().hasLabel('person').as('p').out('created').group().by('name').by(select('p').values('age').sum())
Local traversals may not traverse past the local star-graph on GraphComputer:
[SelectOneStep(last,p), PropertiesStep([age],value), SumGlobalStep]
```
`docker/build.sh -t -i -n` passed.
VOTE +1
[ Full content available at: https://github.com/apache/tinkerpop/pull/997 ]
This message was relayed via gitbox.apache.org for dev@tinkerpop.apache.org
