fix: avoid sourceId collision for messages without Message-ID (CM-1318)

Signed-off-by: Uroš Marolt <uros@marolt.me>

Author: themarolt

services/apps/mailing_list_integration/src/crowdmail/services/parse/noteren.py

@@ -268,8 +268,10 @@ def parse_email(
                     git_id[:12],
                 )
 
-    # Grab the message id for this message
-    msgid = message_id_cleanup(e.get("Message-ID", None))
+    # Grab the message id for this message. Fall back to a synthetic id built
+    # from the (unique) git commit sha when Message-ID is absent, so the
+    # sourceId-based dedup key can't collide across such messages.
+    msgid = message_id_cleanup(e.get("Message-ID", None)) or f"{channel}:{git_id}"
 
     # Grab only the "last" reference in the list of references for this email.
     # "References:" is a long list, trying to show everything else in the

services/apps/mailing_list_integration/src/test/test_noteren.py

@@ -736,11 +736,11 @@ def test_run_command_returns_nonzero_exit_code():
     assert rc == 42
 
 
-def test_parse_email_no_message_id_gives_empty_source_id():
-    """A message without a Message-ID header must yield sourceId == ''."""
+def test_parse_email_no_message_id_gives_synthetic_source_id():
+    """A message without a Message-ID header must fall back to a channel:git_id sourceId."""
     raw = b"From: A <a@example.com>\nSubject: s\nDate: Mon, 1 Jan 2024 12:00:00 +0000\n\nbody\n"
     parsed = _parse_email(raw, "src", "chan", "c1", "b1")
-    assert parsed["activityData"]["sourceId"] == ""
+    assert parsed["activityData"]["sourceId"] == "chan:c1"
 
 
 def test_parse_email_no_subject_gives_none_title():
@@ -856,10 +856,10 @@ def test_parse_email_url_contains_message_id():
 
 
 def test_parse_email_url_with_no_message_id():
-    """When Message-ID is absent the url must end with a slash under the source's /r/ path."""
+    """When Message-ID is absent the url must embed the synthetic channel:git_id id."""
     raw = b"From: A <a@example.com>\nSubject: s\nDate: Mon, 1 Jan 2024 12:00:00 +0000\n\nbody\n"
     parsed = _parse_email(raw, "src", "chan", "c1", "b1")
-    assert parsed["activityData"]["url"] == "src/r/"
+    assert parsed["activityData"]["url"] == "src/r/chan:c1"
 
 
 def test_parse_email_result_structure():