Adds support for emojis and more IDNA URLs (#161)

* Adds support for emojis and more IDNA URLs
* Validates URLs with IPs ending in .0 or .255


This reverts commit 42ea2fd.

Author: automationator

tests/test_url.py

@@ -42,12 +42,18 @@
     u'http://مثال.إختبار',
     u'http://例子.测试',
     u'http://उदाहरण.परीक्षा',
+    u'http://www.😉.com',
+    u'http://😉.com/😁',
+    u'http://উদাহরণ.বাংলা',
+    u'http://xn--d5b6ci4b4b3a.xn--54b7fta0cc',
     u'http://-.~_!$&\'()*+,;=:%40:80%2f::::::@example.com',
     u'http://1337.net',
     u'http://a.b-c.de',
     u'http://223.255.255.254',
+    u'http://10.1.1.0',
     u'http://10.1.1.1',
     u'http://10.1.1.254',
+    u'http://10.1.1.255',
     u'http://127.0.0.1:8080',
     u'http://127.0.10.150',
     u'http://localhost',
@@ -114,8 +120,6 @@ def test_returns_true_on_valid_public_url(address, public):
     'http://-a.b.co',
     'http://a.b-.co',
     'http://0.0.0.0',
-    'http://10.1.1.0',
-    'http://10.1.1.255',
     'http://224.1.1.1',
     'http://1.1.1.1.1',
     'http://123.123.123',

validators/url.py

@@ -3,7 +3,7 @@
 from .utils import validator
 
 ip_middle_octet = r"(?:\.(?:1?\d{1,2}|2[0-4]\d|25[0-5]))"
-ip_last_octet = r"(?:\.(?:[1-9]\d?|1\d\d|2[0-4]\d|25[0-4]))"
+ip_last_octet = r"(?:\.(?:0|[1-9]\d?|1\d\d|2[0-4]\d|25[0-5]))"
 
 regex = re.compile(  # noqa: W605
     r"^"
@@ -69,16 +69,19 @@
     r"(25[0-5]|(2[0-4]|1{0,1}[0-9]){0,1}[0-9])"
     r")\]|"
     # host name
-    r"(?:(?:[a-z\u00a1-\uffff0-9]-?)*[a-z\u00a1-\uffff0-9]+)"
+    r"(?:(?:(?:xn--)|[a-z\u00a1-\uffff\U00010000-\U0010ffff0-9]-?)*"
+    r"[a-z\u00a1-\uffff\U00010000-\U0010ffff0-9]+)"
     # domain name
-    r"(?:\.(?:[a-z\u00a1-\uffff0-9]-?)*[a-z\u00a1-\uffff0-9]+)*"
+    r"(?:\.(?:(?:xn--)|[a-z\u00a1-\uffff\U00010000-\U0010ffff0-9]-?)*"
+    r"[a-z\u00a1-\uffff\U00010000-\U0010ffff0-9]+)*"
     # TLD identifier
-    r"(?:\.(?:[a-z\u00a1-\uffff]{2,}))"
+    r"(?:\.(?:(?:xn--[a-z\u00a1-\uffff\U00010000-\U0010ffff0-9]{2,})|"
+    r"[a-z\u00a1-\uffff\U00010000-\U0010ffff]{2,}))"
     r")"
     # port number
     r"(?::\d{2,5})?"
     # resource path
-    u"(?:/[-a-z\u00a1-\uffff0-9._~%!$&'()*+,;=:@/]*)?"
+    u"(?:/[-a-z\u00a1-\uffff\U00010000-\U0010ffff0-9._~%!$&'()*+,;=:@/]*)?"
     # query string
     r"(?:\?\S*)?"
     # fragment