Accept only 4 decimal octet IPv4 addresses. Support IPv4 addresses wi…

…th unicode dots. (#292)

- IPv4 addresses with unicode dots are now recognized. Closes #287
- IPv4 addresses must have 4 decimal octets. Closes #290

---------

Co-authored-by: John Kurkowski <[email protected]>

tests/main_test.py

@@ -13,6 +13,7 @@
 import tldextract
 import tldextract.suffix_list
 from tldextract.cache import DiskCache
+from tldextract.remote import inet_pton, looks_like_ip
 from tldextract.suffix_list import SuffixListNotFound
 from tldextract.tldextract import ExtractResult
 
@@ -132,7 +133,18 @@ def test_ip():
     )
 
 
-def test_looks_like_ip():
+@pytest.mark.skipif(not inet_pton, reason="inet_pton unavailable")
+def test_looks_like_ip_with_inet_pton():
+    assert looks_like_ip("1.1.1.1", inet_pton) is True
+    assert looks_like_ip("256.256.256.256", inet_pton) is False
+
+
+def test_looks_like_ip_without_inet_pton():
+    assert looks_like_ip("1.1.1.1", None) is True
+    assert looks_like_ip("256.256.256.256", None) is False
+
+
+def test_similar_to_ip():
     assert_extract("1\xe9", ("", "", "1\xe9", ""))
 
 
@@ -328,6 +340,11 @@ def test_ipv4():
         ("", "", "127.0.0.1", ""),
         expected_ip_data="127.0.0.1",
     )
+    assert_extract(
+        "http://127\u30020\uff0e0\uff611/foo/bar",
+        ("", "", "127.0.0.1", ""),
+        expected_ip_data="127.0.0.1",
+    )
 
 
 def test_ipv4_bad():
@@ -339,6 +356,12 @@ def test_ipv4_bad():
 
 
 def test_ipv4_lookalike():
+    assert_extract(
+        "http://127.0.0/foo/bar", ("", "127.0", "0", ""), expected_ip_data=""
+    )
+    assert_extract(
+        "http://127.0.0.0x1/foo/bar", ("", "127.0.0", "0x1", ""), expected_ip_data=""
+    )
     assert_extract(
         "http://127.0.0.1.9/foo/bar", ("", "127.0.0.1", "9", ""), expected_ip_data=""
     )

tldextract/remote.py

@@ -1,9 +1,17 @@
 """tldextract helpers for testing and fetching remote resources."""
 
+from __future__ import annotations
+
 import re
-import socket
+from collections.abc import Callable
 from urllib.parse import scheme_chars
 
+inet_pton: Callable[[int, str], bytes] | None
+try:
+    from socket import AF_INET, inet_pton  # Availability: Unix, Windows.
+except ImportError:
+    inet_pton = None
+
 IP_RE = re.compile(
     r"^(([0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|25[0-5])\.)"
     r"{3}([0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|25[0-5])$"
@@ -44,18 +52,17 @@ def _schemeless_url(url: str) -> str:
     return url[double_slashes_start + 2 :]
 
 
-def looks_like_ip(maybe_ip: str) -> bool:
+def looks_like_ip(
+    maybe_ip: str, pton: Callable[[int, str], bytes] | None = inet_pton
+) -> bool:
     """Check whether the given str looks like an IP address."""
     if not maybe_ip[0].isdigit():
         return False
 
-    try:
-        socket.inet_aton(maybe_ip)
-        return True
-    except (AttributeError, UnicodeError):
-        if IP_RE.match(maybe_ip):
+    if pton is not None:
+        try:
+            pton(AF_INET, maybe_ip)
             return True
-    except OSError:
-        pass
-
-    return False
+        except OSError:
+            return False
+    return IP_RE.match(maybe_ip) is not None

tldextract/tldextract.py

@@ -251,19 +251,19 @@ def extract_urllib(
     def _extract_netloc(
         self, netloc: str, include_psl_private_domains: bool | None
     ) -> ExtractResult:
-        labels = (
+        netloc_with_ascii_dots = (
             netloc.replace("\u3002", "\u002e")
             .replace("\uff0e", "\u002e")
             .replace("\uff61", "\u002e")
-            .split(".")
         )
+        labels = netloc_with_ascii_dots.split(".")
 
         suffix_index = self._get_tld_extractor().suffix_index(
             labels, include_psl_private_domains=include_psl_private_domains
         )
 
-        if suffix_index == len(labels) and netloc and looks_like_ip(netloc):
-            return ExtractResult("", netloc, "")
+        if suffix_index == len(labels) == 4 and looks_like_ip(netloc_with_ascii_dots):
+            return ExtractResult("", netloc_with_ascii_dots, "")
 
         suffix = ".".join(labels[suffix_index:]) if suffix_index != len(labels) else ""
         subdomain = ".".join(labels[: suffix_index - 1]) if suffix_index >= 2 else ""