Merge pull request #243 from oasis-open/claude-fix-180

Fix for issue #180 by Claude Code

Author: ejratl

docs/best-practices.rst

@@ -99,8 +99,8 @@ Mandatory Checks - STIX 2.1
 |                                 | RFC 5646 language code                 | language code.                         |
 +---------------------------------+----------------------------------------+----------------------------------------+
 | software_language               | the 'language' property of software    | The 'languages' property of object     |
-|                                 | objects is a valid ISO 639-2 language  | '<identifier>' contains an invalid     |
-|                                 | code                                   | code ('<lang>').                       |
+|                                 | objects is a valid RFC 5646 language   | '<identifier>' contains an invalid     |
+|                                 | code (ISO 639-2 accepted with warning) | code ('<lang>').                       |
 +---------------------------------+----------------------------------------+----------------------------------------+
 | patterns                        | that the syntax of the pattern of an   | '<object>' is not a valid observable   |
 |                                 | indicator is valid, and that objects   | type name                              |

pinning-schema.md

@@ -12,6 +12,6 @@ Output confirms the schema 2.1 correct commit is pinned:
 Entering 'stix2validator/schemas-2.0'
 stix2validator/schemas-2.0 b155093705ab4934ee29e7ba4dc99ed053cd4e7f
 Entering 'stix2validator/schemas-2.1'
-stix2validator/schemas-2.1 4d010b385b83e53ea322357783ff3be3da5820fb
+stix2validator/schemas-2.1 c4f8d589acf2bdb3783655c89e0ffb6e150006ae
 
-Push from local to my repo (fork of Oasis) which then feeds the change to the PR on Oasis.  I've updated the comments in the PR to confirm.
+Push from local to my repo (fork of Oasis) which then feeds the change to the PR on Oasis.  I've updated the comments in the PR to confirm.

stix2validator/schemas-2.1

@@ -643,9 +643,41 @@ def test_software_language(self):
         }
         self.assertFalseWithOptions(observed_data)
 
-        observed_data['languages'][0] = 'eng'
+        # Test with RFC 5646 language code (preferred)
+        observed_data['languages'][0] = 'en'
         self.assertTrueWithOptions(observed_data)
 
+    def test_software_language_iso639_compatibility(self):
+        """Test that ISO 639-2 codes are accepted with warnings for backward compatibility"""
+        import io
+        import logging
+
+        # Capture log output
+        log_capture = io.StringIO()
+        handler = logging.StreamHandler(log_capture)
+        handler.setLevel(logging.WARNING)
+        logger = logging.getLogger('stix2validator.v21.musts')
+        logger.addHandler(handler)
+        logger.setLevel(logging.WARNING)
+
+        observed_data = {
+            "type": "software",
+            "id": "software--ff1e0780-358c-5808-a8c7-d0fca4ef6ef4",
+            "name": "word",
+            "languages": ["eng"]  # ISO 639-2 code
+        }
+
+        # Should be valid (backward compatibility)
+        self.assertTrueWithOptions(observed_data)
+
+        # Should generate a warning
+        log_output = log_capture.getvalue()
+        self.assertIn("ISO 639-2 language code", log_output)
+        self.assertIn("RFC 5646 language codes are preferred", log_output)
+
+        # Clean up
+        logger.removeHandler(handler)
+
     def test_software_cpe(self):
         observed_data = {
             "type": "software",

stix2validator/test/v21/observed_data_tests.py

@@ -411,16 +411,30 @@ def language(instance):
 
 @cyber_observable_check("2.1")
 def software_language(instance):
-    """Ensure the 'language' property of software objects is a valid ISO 639-2
-    language code.
+    """Ensure the 'language' property of software objects is a valid RFC 5646
+    language code. ISO 639-2 codes are accepted for backward compatibility
+    but will generate warnings.
     """
     if ('type' in instance and instance['type'] == 'software' and
             'languages' in instance):
         for lang in instance['languages']:
-            if lang not in enums.SOFTWARE_LANG_CODES:
+            if lang in enums.LANG_CODES:
+                # Valid RFC 5646 language code
+                continue
+            elif lang in enums.SOFTWARE_LANG_CODES:
+                # ISO 639-2 code - accept but warn
+                import logging
+                logger = logging.getLogger(__name__)
+                logger.warning("The 'languages' property of object '%s' "
+                               "contains an ISO 639-2 language code ('%s'). "
+                               "RFC 5646 language codes are preferred for STIX 2.1. "
+                               "Consider updating to RFC 5646 format."
+                               % (instance['id'], lang))
+            else:
+                # Invalid language code
                 yield JSONError("The 'languages' property of object '%s' "
-                                "contains an invalid ISO 639-2 language "
-                                " code ('%s')."
+                                "contains an invalid language code ('%s'). "
+                                "Expected RFC 5646 language code."
                                 % (instance['id'], lang), instance['id'])