Incorrect empty FQDN when no subdomain and domain

[BookGin]

When the extract result's subdomain and domain are both empty, it will return an incorrect empty FQDN.

Python 3.6.7 (default, Oct 22 2018, 11:32:17) 
[GCC 8.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import tldextract
>>> tldextract.__version__
'2.2.0'
>>> tldextract.extract('http://s3.amazonaws.com/')
ExtractResult(subdomain='', domain='', suffix='s3.amazonaws.com')
>>> tldextract.extract('http://s3.amazonaws.com/').fqdn
''
>>> tldextract.extract('http://blogspot.com/')
ExtractResult(subdomain='', domain='', suffix='blogspot.com')
>>> tldextract.extract('http://blogspot.com/').fqdn
''

#174 would be a related issue but it's not the same.

Is this check really necessary?

tldextract/tldextract/tldextract.py

Lines 121 to 124 in d657414

	if self.domain and self.suffix:
	# self is the namedtuple (subdomain domain suffix)
	return '.'.join(i for i in self if i)
	return ''

[john-kurkowski]

Ugh, that is unintuitive. Private domains strike again. The check was originally for e.g. tldextract.extract('localhost'), where there is no FQDN to reconstruct. It didn't consider private domains.

Like this comment in #138, if we add a way for each suffix to know whether it is private or not, we could change the check to something like this.

if self.suffix and (self.domain or self.is_private):