Performance enhancements related to issue #772

[benlamonica]

Unfortunately I don't have the time to continue pursuing the solution to this problem, but I thought I would send the small code changes that I made that improved performance in this extreme situation.

[benlamonica]

Removed the hashCode/equals change, as it broke tests.

[cbeust]

@juherr Thoughts on this PR?

[juherr]

@cbeust I never feel good perf improvments without something concrete to evaluate it.

@benlamonica Could you add a test based on your ContentionTest from #772? How many time should it take before and after the fix? We can imagine the new test should fail if it takes too many time.

[cbeust]

@juherr Agreed.

Unfortunately, @benlamonica says that he can't work on this any more so it's up to us to follow up or close this, and I agree with you that I wouldn't feel comfortable merging this until we can run some numbers and prove that it improves performance significantly.

[juherr]

@cbeust I looked further and the test case is interesting. It showed me that it has many problems but I don't know how to fix them for the moment. I suppose it will involve many deep changes.

For the current PR, I've no idea what to do. We just know it doesn't broke something (for the moment) and it helps @benlamonica. But it is clear we should keep #772 open.