I agree with reversion. This is an attempt at low level performance optimisation, and that is a task better suited to the JVM and Hotspot.
It is highly unfortunate that tools have ever highlighted this as being an 'issue' that needs 'fixing'. All the advice I have been hearing recently, and it corresponds to how I now write code, is that you should write the correctly designed code, and just allow Hotspot to sort out issues. Only if you find a specific problem should performance tweaks like this ever be considered. Stephen ----- Original Message ---- From: Phil Steitz <[EMAIL PROTECTED]> Thanks for the explanation. I would, however, like to ask that this change be reverted and a JIRA ticket be opened to evaluate this API change. If real performance improvements can be demonstrated, we can consider either exposing this private inner class or removing it. Phil --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
