class SplittableRandom

David Chase at
Fri Jul 19 20:14:56 UTC 2013

On 2013-07-19, at 2:00 PM, Mike Duigou <mike.duigou at> wrote:
> 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. Solutions which use lousy seeding until SecureRandom is available won't be acceptable.

It seems to have a fallback implementation when no security providers are available.  Is that acceptably random on most platforms?

> A mechanism to get entropy very early in VM boot would definitely be useful.

Ideally, configurable entropy that could be recorded and replayed for debugging.


More information about the core-libs-dev mailing list