JDK 8 RFR 6470700: Math.random() / Math.initRNG() uses double checked locking

2013-08-22 Thread Brian Burkhalter
JDK 8 Reviewers: The proposed patch is here: http://cr.openjdk.java.net/~bpb/6470700/ JTREG tests pass and JMH benchmarking shows a very slight performance improvement versus the extant code. Thanks, Brian On Aug 21, 2013, at 7:23 PM, Mike Duigou wrote: On Aug 21 2013, at 17:01 , David M.

Re: JDK 8 RFR 6470700: Math.random() / Math.initRNG() uses double checked locking

2013-08-22 Thread Alan Bateman
On 22/08/2013 20:41, Brian Burkhalter wrote: JDK 8 Reviewers: The proposed patch is here: http://cr.openjdk.java.net/~bpb/6470700/ JTREG tests pass and JMH benchmarking shows a very slight performance improvement versus the extant code. It looks okay to me although the downside of this

Re: JDK 8 RFR 6470700: Math.random() / Math.initRNG() uses double checked locking

2013-08-22 Thread Brian Burkhalter
You are correct: it should be private - I'll fix that. Thanks, Brian On Aug 22, 2013, at 12:46 PM, Alan Bateman wrote: It looks okay to me although the downside of this idiom is yet another class (which I assume can private btw).

Re: JDK 8 RFR 6470700: Math.random() / Math.initRNG() uses double checked locking

2013-08-22 Thread Chris Hegarty
On 08/22/2013 08:46 PM, Alan Bateman wrote: On 22/08/2013 20:41, Brian Burkhalter wrote: JDK 8 Reviewers: The proposed patch is here: http://cr.openjdk.java.net/~bpb/6470700/ JTREG tests pass and JMH benchmarking shows a very slight performance improvement versus the extant code. It looks