(Catching up...) On 07/19/13 14:00, Mike Duigou wrote:
Using a shared static SecureRandom.generateSeed() to generate the seed is probably the easiest way to get a good seed that is portable.
I agree and updated accordingly.
I have previous experience mucking in this area. Attempting to use SecureRandom will likely encounter order of initialization issues as it is loaded from security providers and is only functional/available very late in the VM boot process.
Because SplittableRandom is a new class, there won't be any immediate within-JDK dependencies, so this shouldn't be a problem.
I had considered suggesting adding a vm interface to access the platform native secure random number generation (/dev/random or rand_s). This would be useful for early boot and perhaps for speeding up seeding of SecureRandom.
Good idea. I started to explore doing something like this only for SplittableRandom, but then noticed that it would be better to benefit from more general improvements to SecureRandom in using OS-supported randoms: When OS support exists (which is almost always these days), it could and should be made simpler and faster to use. (Additionally, a "getLongSeed()" method would be handy, avoiding two byte-array conversions in what is probably the most common use case.) -Doug
