FWIW, I've pinged my brother on this question; since he was in Sun's Java
performance team he should have A Clue about proper practice now and in
the future. As I say, my understanding based on past discussion with him
agrees with Chris's -- the presumption that final has performance effects
is idiosyncratic to particular JVMs, not guaranteed, and not recommended.
It's possible that things have changed or I misunderstood, which is why
I'm cross-checking.
I'm a firm believer that if you're having to do something like this to
trick the compiler into doing what you want, you should instead be
reporting it to the compiler's authors and getting them to improve the
optimizer.
______________________________________
Joe Kesselman, IBM Next-Generation Web Technologies: XML, XSL and more.
"may'ron DaroQbe'chugh vaj bIrIQbej" ("Put down the squeezebox and nobody
gets hurt.")
