Anton Hägerstrand

I benchmarked the current commit (`e826696fd235cb1ef215745944cac6e34faa0d20`) using the same modified lucenebench as above, it seems to like this change even better: ``` TaskQPS baseline StdDevQPS my_modified_version StdDev Pct diff p-value...

> Before merging I'd be curious to better understand why the JVM doesn't optimize this better. Presumably, it should be able to resolve the virtual call once for the entire...

> Ideally we'd add a query that adds another implementation of an `IntersectVisitor` to the nightly benchmarks before merging that PR so that we can see the performance bump? Yes...

Thank you all for helping to get this PR into a better state. The only thing that still irked me is that I couldn't get the tests to execute all...

> One of we committers must merge it! It sounds like we are super close ... I'll try to review today and maybe merge. Sounds great - the javadoc fix...

> Thanks @antonha -- sorry, maybe also add a `CHANGES.txt` entry? This is an exciting opto! My bad - I should have added one the first time you asked :see_no_evil:....

After banging my head against this issue for way too long, I have a theory. This is running play 2.6.7, but I think it probably applies to earlier versions as...

Reading up on the original reporter's issue, I think that my 'theory' was for a different bug than what was originally reported, but with the same symptoms. Might still be...