It took a few more minor tweaks, but I think I finally have TinkerFactory "right". Here's the final look (i hope):
https://github.com/apache/tinkerpop/blob/57e7f702e44356b72ffa7d1d11d526 ac5e768812/tinkergraph-gremlin/src/main/java/org/apache/tinkerpop/gremlin/ tinkergraph/structure/TinkerFactory.java I ended up reverting createClassic() to how it was working before - it did some weird stuff to some of the integration tests for some reason and I wasn't sure why, so seemed best to keep it as it was. That method was already "mostly" right - I'm just not sure why it didn't use a long IdManager for vertex properties, but that's probably fine. On Fri, Jun 2, 2017 at 7:41 PM, Stephen Mallette <[email protected]> wrote: > This really surprised me: > > gremlin> graph = TinkerFactory.createModern() > ==>tinkergraph[vertices:6 edges:6] > gremlin> g = graph.traversal() > ==>graphtraversalsource[tinkergraph[vertices:6 edges:6], standard] > gremlin> g.V(1) > ==>v[1] > gremlin> g.V(1L) > gremlin> > > We all know why it does that - I'm just amazed that createModern() isn't > rigged with an IdManager to handle it properly. createClassic() is just > fine: > > gremlin> graph = TinkerFactory.createClassic() > ==>tinkergraph[vertices:6 edges:6] > gremlin> g = graph.traversal() > ==>graphtraversalsource[tinkergraph[vertices:6 edges:6], standard] > gremlin> g.V(1) > ==>v[1] > gremlin> g.V(1L) > ==>v[1] > > Going to push a fix shortly to tp31 on up (can't seem to think of a single > reason it is setup this way - so weird). I'll then issue the code freeze > email to follow. > > >
