Base64-decoded secrets should not contain strictly ANSI alphanumeric, etc. data

[michaelcfanning]

This very likely indicates a false positive, eg., we have uncovered a base64-encoded URL that is provided as a query string parameter in an HTTP request (a real world case we encountered).

We should create a helper and call it during the static analysis phase for all relevant rules, returning 'no match' in cases where conditions aren't met (i.e., detecting one or more characters outside the latin unicode range).

[michaelcfanning]

@eddynaka