Log random random seed

One can argue that in many cases one should not set a specific random seed. Instead just use a random random seed and log that, so that one could set it explicitly later on when reproducability is an issue, c.f. https://blog.genesmindsmachines.com/p/random-seeds-and-brown-m-and-ms. In the case of `dqrng` that would mean logging the seed obtained from R's RNG at start-up.
