making fix_dir_mtime.py more robust

fix_dir_mtime.py

@@ -1,34 +1,71 @@
 #!/usr/bin/env python3
-
 import os
 import sys
 from pathlib import Path
 
 
-def update_dir_mtime(path: Path):
+def stat_follow(p: Path):
+    """stat() that follows symlinks; returns None if target missing."""
+    try:
+        return p.stat()  # follows symlinks
+    except FileNotFoundError:
+        return None
+
+
+def update_dir_mtime(path: Path, visited: set[tuple[int, int]]) -> float:
+    """
+    Make directory mtime >= newest reachable content mtime.
+    Follows working symlinks; ignores broken ones.
+    Prevents cycles via inode tracking.
+    Returns the newest mtime found under 'path' (including itself after update).
+    """
+    st = stat_follow(path)
+    if st is None:
+        return 0.0
+
+    key = (st.st_dev, st.st_ino)
+    if key in visited:
+        return st.st_mtime  # cycle: stop here
+    visited.add(key)
+
+    # If it's not a directory, just return its mtime.
     if not path.is_dir():
-        return path.stat().st_mtime
+        return st.st_mtime
 
-    newest = 0
+    newest = 0.0
 
-    for child in path.iterdir():
-        if child.is_dir():
-            child_mtime = update_dir_mtime(child)
-        else:
-            child_mtime = child.stat().st_mtime
+    try:
+        for child in path.iterdir():
+            # Follow working symlinks (child.stat() will fail if broken).
+            child_st = stat_follow(child)
+            if child_st is None:
+                continue  # broken symlink or vanished entry -> ignore
 
-        newest = max(newest, child_mtime)
+            if (
+                child.is_dir()
+            ):  # this is True for symlink->dir as well (when target exists)
+                child_mtime = update_dir_mtime(child, visited)
+            else:
+                child_mtime = child_st.st_mtime
 
-    if newest > 0:
+            if child_mtime > newest:
+                newest = child_mtime
+    except PermissionError:
+        # If you hit protected folders, skip descending but still allow returning current mtime.
+        newest = 0.0
+
+    # Update this directory to reflect newest reachable content.
+    if newest > 0.0 and newest > st.st_mtime:
         os.utime(path, (newest, newest))
+        st = stat_follow(path) or st
 
-    return max(newest, path.stat().st_mtime)
+    return max(newest, st.st_mtime)
 
 
 if __name__ == "__main__":
     if len(sys.argv) != 2:
         print("Usage: fix_dir_mtime.py <directory>")
-        sys.exit(1)
+        raise SystemExit(1)
 
     root = Path(sys.argv[1]).resolve()
-    update_dir_mtime(root)
+    update_dir_mtime(root, visited=set())