test_hostname_regex

#!/usr/bin/env python
# vim: set fileencoding=utf=8 :
'''Test regular expressions for matching an Internet host name

These regular expressions were found at
http://stackoverflow.com/questions/1418423/the-hostname-regex
'''
from __future__ import unicode_literals
import re


regexes = [
    re.compile( # http://stackoverflow.com/q/1418423/187736
        r'^[0-9a-z]([0-9a-z\-]{0,61}[0-9a-z])?'
        r'(\.[0-9a-z](0-9a-z\-]{0,61}[0-9a-z])?)*$'
    ),
    re.compile( # http://stackoverflow.com/a/1420225/187736
        r'^(?=.{1,255}$)[0-9A-Za-z]'
        r'(?:(?:[0-9A-Za-z]|-){0,61}[0-9A-Za-z])?'
        r'(?:\.[0-9A-Za-z](?:(?:[0-9A-Za-z]|-){0,61}[0-9A-Za-z])?)*\.?$'
    ),
    re.compile( # http://stackoverflow.com/a/1418724/187736
        r'(?:(?:(?:(?:[a-zA-Z0-9][-a-zA-Z0-9]*)?[a-zA-Z0-9])[.])*'
        r'(?:[a-zA-Z][-a-zA-Z0-9]*[a-zA-Z0-9]|[a-zA-Z])[.]?)'
    ),
    re.compile( # http://stackoverflow.com/a/1418724/187736
        r'(?:(?:(?:(?:[a-zA-Z0-9][-a-zA-Z0-9]{0,61})?[a-zA-Z0-9])[.])*'
        r'(?:[a-zA-Z][-a-zA-Z0-9]{0,61}[a-zA-Z0-9]|[a-zA-Z])[.]?)'
    ),
    # This one doesn't even compile
    #re.compile( # http://stackoverflow.com/a/18494710/187736
    #    r'^([a-zA-Z0-9](?:(?:[a-zA-Z0-9-]*|(?<!-)\.(?![-.]))*[a-zA-Z0-9]+)?)$'
    #),
    re.compile( # http://stackoverflow.com/a/10874552/187736
        r'^(?=.{1,255})([0-9A-Za-z]|_{1}|\*{1}$)'
        r'(?:(?:[0-9A-Za-z]|\b-){0,61}[0-9A-Za-z])?'
        r'(?:\.[0-9A-Za-z](?:(?:[0-9A-Za-z]|\b-){0,61}[0-9A-Za-z])?)*\.?$'
    ),
]
regex = regexes[1]


def test_single_label():
    '''ensure the regex matches a single-label name'''

    assert regex.match('hostname') is not None


def test_local():
    '''ensure the regex matches a two-part name'''

    assert regex.match('hostname.local') is not None


def test_fqdn():
    '''ensure the regex matches a three-part name'''

    assert regex.match('hostname.domain.tld') is not None


def test_fqdn_leaf():
    '''ensure the regex matches a four-part name'''

    assert regex.match('leaf.hostname.domain.tld') is not None


def check_invalid(name):
    '''ensure the regex does not match a name containing {0}'''
    assert regex.match(name) is None


def test_invalid():
    '''ensure the regex does not match a name with invalid characters'''

    syms = '`~!@#$%^&*()+=_'
    for s in syms:
        check_invalid.description = check_invalid.__doc__.format(s)
        yield check_invalid, s


def test_whitespace():
    '''ensure the regex does not match a name containing whitespace'''

    assert regex.match('host name') is None


def test_hyphen():
    '''ensure the regex matches a name containing a hyphen'''

    assert regex.match('host-name') is not None

def test_leading_hyphen():
    '''ensure the regex does not match a name starting with a hyphen'''

    assert regex.match('-hostname') is None


def test_trailing_hyphen():
    '''ensure the regex does not match a name ending with a hyphen'''

    assert regex.match('hostname-') is None


def test_leading_period():
    '''ensure the regex does not match a name starting with a dot'''

    assert regex.match('.hostname') is None


def test_trailing_period():
    '''ensure the regex matches a name ending with a dot'''

    assert regex.match('hostname.') is not None


def test_only_dot():
    '''ensure the regex does not match only a single dot'''

    assert regex.match('.') is None


def test_max_len_single_label():
    '''ensure the regex does not match a 64-character single name'''

    assert regex.match(64 * 'a') is None


def test_max_len_single_label2():
    '''ensure the regex matches a 63-character single name'''

    assert regex.match(63 * 'a') is not None


def test_max_len_tld():
    '''ensure the regex matches a 63-character top level name'''

    assert regex.match(63 * 'a' + '.') is not None


def test_max_len_fqdn():
    '''ensure the regex matches a 63-character hostname in a FQDN'''

    assert regex.match(63 * 'a' + '.' + 'b') is not None


def test_max_len_fqdn():
    '''ensure the regex matches a 255-character FQDN'''

    h = '.'.join((
        63 * 'a',
        63 * 'b',
        63 * 'c',
        63 * 'd',
    ))
    assert regex.match(h) is not None


def test_max_len_fqdn2():
    '''ensure the regex does not match a 255-character 2-part FQDN'''

    h = '.'.join((
        63 * 'a',
        191 * 'b',
    ))
    assert regex.match(h) is None


def test_double_dot():
    '''ensure the regex does not match multiple successive dots'''

    assert regex.match('hostname..domain') is None


def test_idn():
    '''ensure the regex does not match a UTF-8 encoded IDN'''

    assert regex.match('bücher') is None


def test_idn_punycode():
    '''ensure the regex matches a Punycode encoded IDN'''

    assert regex.match('bücher'.encode('idna').decode('ascii')) is not None