This change makes myparser.emails(...) stricter so that, given a word
(eg: 'word.com') it does not match domains which end with that word
(i.e. some-other-word.com). Subdomains continue to be matched (eg:
'subdomain.word.com').
Also add unit test for emails(...)