Encode hostname directly instead of netloc

tests/test_url.py

@@ -1104,6 +1104,18 @@ def test_scheme_case_insensitive(self):
         result = parse_data_uri("DaTa:,A%20brief%20note")
         self.assertEqual(result.data, b"A brief note")
 
+    def test_safe_url_string_encode_idna_domain_with_port(self):
+        self.assertEqual(safe_url_string('http://新华网.中国:80'),
+                         'http://xn--xkrr14bows.xn--fiqs8s:80')
+
+    def test_safe_url_string_encode_idna_domain_with_username_password_and_port_number(self):
+        self.assertEqual(safe_url_string('ftp://admin:admin@新华网.中国:21'),
+                         'ftp://admin:[email protected]:21')
+
+    def test_safe_url_string_encode_idna_domain_with_username_without_password_and_port_number(self):
+        self.assertEqual(safe_url_string('ftp://admin:@新华网.中国:21'),
+                         'ftp://admin:@xn--xkrr14bows.xn--fiqs8s:21')
+
 
 if __name__ == "__main__":
     unittest.main()

w3lib/url.py

@@ -95,10 +95,26 @@ def safe_url_string(
     decoded = to_unicode(url, encoding=encoding, errors="percentencode")
     parts = urlsplit(_ascii_tab_newline_re.sub("", decoded))
 
+    username, password, hostname, port_number = (
+        parts.username,
+        parts.password,
+        parts.hostname,
+        parts.port
+    )
+    netloc_bytes = b''
+
     # IDNA encoding can fail for too long labels (>63 characters)
     # or missing labels (e.g. http://.example.com)
     try:
-        netloc_bytes = parts.netloc.encode("idna")
+        # When we have hostname we use it instead of netloc directly
+        if hostname:
+            if isinstance(username, str) and isinstance(password, str):
+                netloc_bytes += f'{username}:{password}@'.encode("idna")
+            netloc_bytes += hostname.encode("idna")
+            if port_number:
+                netloc_bytes += f":{port_number}".encode("idna")
+        else:
+            netloc_bytes = parts.netloc.encode("idna")
     except UnicodeError:
         netloc = parts.netloc
     else: