On Thu, Jul 03, 2008 at 07:38:27PM -0700, tim Rowledge wrote:
> The inlining/internalising of the bytecode loop is nasty,  
> resulting in near duplicates of many methods for no very good reason.

It might be nasty, but it's done that way for a good (albeit poorly
documented) reason. The near-duplicate case targets are written that
way to enable early fetching of the next bytecode, which presumably
carries a performance benefit on some machines.

Dave