gh-149819: Fix .pth files not loaded in Python subprocesses

BugBounty Mind · BugBounty Mind · commit 3b8367530dae · 2026-05-15T16:02:58.000Z
After PR gh-149583 (Fix double evaluation of .pth and .site files in venvs), .pth files are no longer loaded in subprocesses started with subprocess.run([sys.executable, ...]). The root cause: main() seeds known_paths from removeduppaths() with all sys.path entries inherited from the parent process. addsitedir() then skips .pth processing for every directory already in known_paths. Fix: - main(): call removeduppaths() for dedup but start known_paths as a fresh empty set, so that addsitedir() processes .pth files in every site-packages directory regardless of inherited sys.path. - addsitedir(): move known_paths.add() before the sys.path.append and guard the append with 'sitedir not in sys.path' to avoid creating duplicate entries when called with a fresh known_paths. This preserves the gh-75723 dedup guarantee while allowing subprocesses to load .pth files.
diff --git a/Lib/site.py b/Lib/site.py
@@ -490,13 +490,16 @@ def addsitedir(sitedir, known_paths=None, *, defer_processing_start_files=False)
         reset = False
     sitedir, sitedircase = makepath(sitedir)
 
-    # If the normcase'd new sitedir isn't already known, append it to
-    # sys.path, keep a record of it, and process all .pth and .start files
-    # found in that directory.  If the new sitedir is known, be sure not
-    # to process all of those more than once!  gh-75723
+    # If the normcase'd new sitedir isn't already known, record it to
+    # prevent re-processing, append it to sys.path (only if not already
+    # present), and process all .pth and .start files found in that
+    # directory.  Use a direct sys.path membership check for the append
+    # guard so that callers (like main()) can pass a fresh known_paths
+    # set while avoiding duplicate sys.path entries (gh-149819).
     if sitedircase not in known_paths:
-        sys.path.append(sitedir)
         known_paths.add(sitedircase)
+        if sitedir not in sys.path:
+            sys.path.append(sitedir)
 
         try:
             names = os.listdir(sitedir)
@@ -1000,7 +1003,8 @@ def main():
     global ENABLE_USER_SITE
 
     orig_path = sys.path[:]
-    known_paths = removeduppaths()
+    removeduppaths()
+    known_paths = set()
     if orig_path != sys.path:
         # removeduppaths() might make sys.path absolute.
         # Fix __file__ of already imported modules too.
