> On Feb 23, 2016, at 3:22 PM, Kim Barrett <kim.barr...@oracle.com> wrote: > > I'm not sure what to do about any of this; I'm just feeling like maybe > the proposed API / implementation isn't quite right yet.
I suggest restore the JavaLangRefAccess used by direct byte buffer code and revisit such explicit mechanism until there is a need. Another concern is that this method gets to run in any arbitrary thread rather than the dedicated reference handler thread / cleaner thread. Mandy