Fix: Skip undecodable TXT records instead of raising exception during SPF lookup

checkdmarc/spf.py

@@ -437,17 +437,22 @@ def query_spf_record(
         )
         spf_record = None
         for record in answers:
-            if record == "Undecodable characters":
-                raise UndecodableCharactersInTXTRecord(
-                    "A TXT record contains undecodable characters."
-                )
             # https://datatracker.ietf.org/doc/html/rfc7208#section-4.5
             #
             # Starting with the set of records that were returned by the lookup,
             # discard records that do not begin with a version section of exactly
             # "v=spf1".  Note that the version section is terminated by either an
             # SP character or the end of the record. As an example, a record with
             # a version section of "v=spf10" does not match and is discarded.
+
+            # Check for undecodable characters
+            if record == "Undecodable characters":
+                # We can't determine if this is an SPF record due to encoding issues
+                warnings.append(
+                    "A TXT record with undecodable characters was skipped."
+                )
+                continue
+
             if record.strip('"').startswith(txt_prefix):
                 spf_txt_records.append(record)
             elif record.startswith(txt_prefix):
@@ -1235,14 +1240,16 @@ def get_spf_record(
         :exc:`checkdmarc.spf.SPFTooManyDNSLookups`
     """
     domain = normalize_domain(domain)
-    record = query_spf_record(
+    query_result = query_spf_record(
         domain,
         nameservers=nameservers,
         resolver=resolver,
         timeout=timeout,
         timeout_retries=timeout_retries,
     )
-    record = record["record"]
+    record = query_result["record"]
+    query_warnings = query_result.get("warnings", [])
+
     parsed_record = parse_spf_record(
         record,
         domain,
@@ -1252,6 +1259,11 @@ def get_spf_record(
         timeout_retries=timeout_retries,
     )
     parsed_record["record"] = record
+
+    # Merge warnings from query_spf_record with warnings from parse_spf_record
+    if query_warnings:
+        parsed_record["warnings"] = query_warnings + parsed_record.get("warnings", [])
+
     return parsed_record
 
 

tests.py

@@ -5,6 +5,7 @@
 
 import os
 import unittest
+from unittest.mock import patch
 
 import checkdmarc
 import checkdmarc.bimi
@@ -561,6 +562,37 @@ def testSPFBrokenExpModifierMacro(self):
             domain,
         )
 
+    def testUndecodableCharactersInNonSPFRecord(self):
+        """Non-SPF TXT records with undecodable characters should be ignored with a warning"""
+        domain = "example.com"
+
+        # Mock query_dns to return:
+        # 1. An undecodable non-SPF TXT record
+        # 2. A valid SPF record
+        with patch('checkdmarc.spf.query_dns') as mock_query_dns:
+            # First call for SPF type records (returns empty)
+            # Second call for TXT records (returns undecodable + valid SPF)
+            mock_query_dns.side_effect = [
+                [],  # No SPF type records
+                [
+                    "Undecodable characters",  # TXT record with undecodable chars
+                    '"v=spf1 include:spf.smtp2go.com -all"'  # Valid SPF record
+                ]
+            ]
+
+            # This should succeed and return the valid SPF record
+            result = checkdmarc.spf.get_spf_record(domain)
+
+            # Verify the SPF record was found
+            self.assertIsNotNone(result["record"])
+            self.assertIn("v=spf1", result["record"])
+
+            # Verify a warning was added for the undecodable record
+            self.assertTrue(len(result["warnings"]) > 0)
+            self.assertTrue(
+                any("TXT record" in w and "undecodable" in w.lower() for w in result["warnings"])
+            )
+
 
 if __name__ == "__main__":
     unittest.main(verbosity=2)