On Tue, Nov 22, 2016 at 4:24 PM, Martin Buchholz <marti...@google.com> wrote:
> > > On Mon, Nov 21, 2016 at 1:06 PM, Paul Sandoz <paul.san...@oracle.com> > wrote: > >> >> See the SpliteratorLateBindingFailFastTest. That could be refactored out >> as it’s conflating two concepts and is too collection/map focused. > > I took a look at that test, and saw that ArrayDeque doesn't get tested there because it's less predictable at throwing CME. ArrayDeque expends fewer resources on catching concurrent modification than classes with a modCount, and that's my own preference as well.