Verified randomness of mapHashToPass character mapping

[irremotus]

The mapping from hash values to password characters should be relatively evenly distributed over the full character set. This needs to be verified somehow, maybe by testing a large number and calculating statistics for the results.