On 4/3/14, 12:46 AM, Mark Thomas wrote:
> On 02/04/2014 22:58, Phil Steitz wrote:
>> It looks to me like v 2 pools are no longer fair in the sense of
>> 1.5.x (modulo lucky threads getting to wait on creates). I agree
>> that the default should be no fairness constraint, but it might be
>> good to
On 02/04/2014 22:58, Phil Steitz wrote:
> It looks to me like v 2 pools are no longer fair in the sense of
> 1.5.x (modulo lucky threads getting to wait on creates). I agree
> that the default should be no fairness constraint, but it might be
> good to make this configurable. This looks doable wi
On 4/2/14, 3:38 PM, Gary Gregory wrote:
> Are the concurrency updates in Java 7 (or 8) of any help here?
The machinery is there in the 1.6 vintage classes we now use. We
just need to make fairness of the lock used by our
LinkedBlockingDeque configurable.
Phil
>
> Gary
>
>
> On Wed, Apr 2, 2014 a
Are the concurrency updates in Java 7 (or 8) of any help here?
Gary
On Wed, Apr 2, 2014 at 5:58 PM, Phil Steitz wrote:
> It looks to me like v 2 pools are no longer fair in the sense of
> 1.5.x (modulo lucky threads getting to wait on creates). I agree
> that the default should be no fairness
It looks to me like v 2 pools are no longer fair in the sense of
1.5.x (modulo lucky threads getting to wait on creates). I agree
that the default should be no fairness constraint, but it might be
good to make this configurable. This looks doable with another
small tweak to LinkedBlockingDeque.
