z230

indexer/config.py

@@ -29,5 +29,5 @@ BACKUP_PATH = os.getenv("BACKUP_PATH")
 # Behaviour
 # =========================
 
-DRY_RUN = os.getenv("DRY_RUN", "1") == "1"
+DRY_RUN = os.getenv("DRY_RUN", "true").lower() in ("1", "true", "yes")
 BATCH_SIZE = int(os.getenv("BATCH_SIZE", 1000))

indexer/db.py

@@ -79,10 +79,15 @@ def batch_insert_files(cur, files_list: list, run_id: int) -> dict:
               f["size"], f["mtime"], f["content_hash"], run_id, run_id)
              for f in chunk]
         )
-        # pymysql executemany: lastrowid = first id in batch
-        first_id = cur.lastrowid
-        for j, f in enumerate(chunk):
-            path_to_id[f["relative_path"]] = first_id + j
+        # Fetch real IDs — lastrowid+j is unreliable with executemany
+        paths = [f["relative_path"] for f in chunk]
+        placeholders = ",".join(["%s"] * len(paths))
+        cur.execute(
+            f"SELECT id, relative_path FROM files WHERE relative_path IN ({placeholders})",
+            paths,
+        )
+        for row in cur.fetchall():
+            path_to_id[row[1]] = row[0]
     return path_to_id
 
 

indexer/scanner.py

@@ -20,11 +20,14 @@ def scan_files(root_path: str) -> dict:
                 continue
             rel_path = os.path.relpath(full_path, root_path).replace("\\", "/")
             rel_dir = os.path.relpath(root, root_path).replace("\\", "/")
+            # Truncate microseconds — MySQL DATETIME rounds to whole seconds,
+            # which causes false "modified" detections on every run.
+            mtime = datetime.fromtimestamp(stat.st_mtime).replace(microsecond=0)
             result[rel_path] = {
                 "full_path": full_path,
                 "file_name": name,
                 "directory": rel_dir,
                 "size": stat.st_size,
-                "mtime": datetime.fromtimestamp(stat.st_mtime),
+                "mtime": mtime,
             }
     return result