Randomized unittests

[Gorge Jingale via Digitalmars-d-learn]

Is there any leverage in the D library for doing randomized unit 
testing? Testing things with a range of possibilities instead of 
fixed. Each time the test is ran a different version is executed. 
This provides more coverage.

How does one create a random number at compile time? Is this the 
only way:

http://dpaste.dzfl.pl/668646ce6d71?