Harden parser for malformed multiline records

tests/test_parser.py

@@ -58,3 +58,19 @@ def test_parse_log_file_tolerates_unterminated_quotes():
     records, _union_keys = parse_log_file(stream)
 
     assert records[0]["msg"] == "broken quoted value"
+
+
+def test_parse_log_file_rebuilds_record_after_embedded_newlines():
+    stream = io.BytesIO(
+        b'v015xxxxdate=2024-02-15 time=09:10:11 msg="hello\n'
+        b'broken-fragment\n'
+        b'world" action=Alert\n'
+        b'v015xxxxdate=2024-02-15 time=09:10:12 msg="next" action=Monitor\n'
+    )
+
+    records, _union_keys = parse_log_file(stream)
+
+    assert len(records) == 2
+    assert records[0]["msg"] == "hellobroken-fragmentworld"
+    assert records[0]["action"] == "Alert"
+    assert records[1]["msg"] == "next"