From 9530f33a6309bd0326039abf9d1c86d929f89968 Mon Sep 17 00:00:00 2001
From: naz <naz@theori.io>
Date: Thu, 13 Jun 2024 16:52:19 +0900
Subject: [PATCH] validate netloc with port number

---
 courlan/filters.py  | 11 +++++++----
 tests/unit_tests.py |  4 ++++
 2 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/courlan/filters.py b/courlan/filters.py
index 51e052c..8158481 100644
--- a/courlan/filters.py
+++ b/courlan/filters.py
@@ -41,7 +41,7 @@
 }
 
 # https://github.com/python-validators/validators/blob/master/src/validators/domain.py
-VALID_DOMAIN = re.compile(
+VALID_DOMAIN_PORT = re.compile(
     # First character of the domain
     r"^(?:[a-zA-Z0-9]"
     # Sub domain + hostname
@@ -49,7 +49,10 @@
     # First 61 characters of the gTLD
     + r"+[A-Za-z0-9][A-Za-z0-9-_]{0,61}"
     # Last character of the gTLD
-    + r"[A-Za-z]$",
+    + r"[A-Za-z]"
+    # Port number
+    + r"(\:(6553[0-5]|655[0-2][0-9]|65[0-4][0-9]{2}|"
+    + r"6[0-4][0-9]{3}|[1-5][0-9]{4}|[1-9][0-9]{0,3}))?$",
     re.IGNORECASE,
 )
 
@@ -151,9 +154,9 @@ def domain_filter(domain: str) -> bool:
         return True
 
     # malformed domains
-    if not VALID_DOMAIN.match(domain):
+    if not VALID_DOMAIN_PORT.match(domain):
         try:
-            if not VALID_DOMAIN.match(domain.encode("idna").decode("utf-8")):
+            if not VALID_DOMAIN_PORT.match(domain.encode("idna").decode("utf-8")):
                 return False
         except UnicodeError:
             return False
diff --git a/tests/unit_tests.py b/tests/unit_tests.py
index 307fa43..4bc2557 100644
--- a/tests/unit_tests.py
+++ b/tests/unit_tests.py
@@ -706,6 +706,10 @@ def test_urlcheck():
     assert check_url("http://[2001:0db8:85a3:0000:0000:8a2e:0370:7334]") is None
     assert check_url("http://1:2:3:4:5:6:7:8:9") is None
 
+    # port
+    assert check_url("http://example.com:80") is not None
+    assert check_url("http://example.com:80:80") is None
+
 
 def test_domain_filter():
     "Test filters related to domain and hostnames."