Traversals reference vertices after deletion

[SmedbergM]

In a multiprocessor environment, a traversal will sometimes reference vertices which another thread/processor has already deleted. This causes the entire traversal to fail in an unrecoverable fashion and throw an uncaught exception.

MWE here. Gist of the output log of one run here.

To be entirely honest, I'm very doubtful that this is a gremlin-scala issue. It seems much more likely that the responsibility lies somewhere in the Java implementation (either in org.neo4j.??? or one of the Apache/Tinkerpop impls). I'd welcome a PR from anyone to my MWE project which peels off the gremlin-scala layer while preserving the observed error.

[mpollmeier]

agreed, I doubt that it's a gremlin-scala issue :)

[SmedbergM]

Can close: 46ddfdd6 removes the dependency on gremlin-scala.

NB: In the course of this, I discovered that Transaction is not thread-safe (in fact it's implemented with Apache's ThreadLocal class). That wasn't the underlying cause of the bug, but it's still worth putting into the documentation somewhere I think.

[mpollmeier]

thank you, will factor that in.