gh-150880: Normalize paths on Windows before appending wildcard (GH-152906)

zainnadeem786 · web-flow · commit 1b4135a2c69f · 2026-07-03T16:30:02.000+01:00
This ensures that we don't turn a "valid" path with trailing spaces into an invalid path with embedded spaces.
diff --git a/Lib/test/test_os/test_os.py b/Lib/test/test_os/test_os.py
@@ -5058,6 +5058,30 @@ def test_current_directory(self):
         finally:
             os.chdir(old_dir)
 
+    @unittest.skipIf(sys.platform != 'win32', "Win32 specific test")
+    def test_windows_trailing_space_path(self):
+        import pathlib
+
+        filename = self.create_file("file.txt")
+        path = self.path + " "
+
+        self.assertTrue(os.path.exists(path))
+        os.stat(path)
+        with open(filename + " ", "rb") as file:
+            self.assertEqual(file.read(), b"python")
+
+        self.assertEqual(os.listdir(path), ["file.txt"])
+        with os.scandir(path) as entries:
+            self.assertEqual([entry.name for entry in entries], ["file.txt"])
+        pathlib_entries = list(pathlib.Path(path).iterdir())
+        self.assertEqual([entry.name for entry in pathlib_entries], ["file.txt"])
+        del pathlib_entries
+
+        extended_path = "\\\\?\\" + path
+        self.assertFalse(os.path.exists(extended_path))
+        self.assertRaises(FileNotFoundError, os.listdir, extended_path)
+        self.assertRaises(FileNotFoundError, os.scandir, extended_path)
+
     def test_repr(self):
         entry = self.create_file_entry()
         self.assertEqual(repr(entry), "<DirEntry 'file.txt'>")
diff --git a/Misc/NEWS.d/next/Library/2026-07-02-22-19-53.gh-issue-150880.vPZ7jK.rst b/Misc/NEWS.d/next/Library/2026-07-02-22-19-53.gh-issue-150880.vPZ7jK.rst
@@ -0,0 +1,3 @@
+Normalize non-extended Windows paths before appending the wildcard used by
+``os.listdir()`` and ``os.scandir()``, making paths with trailing spaces
+behave consistently with other filesystem APIs.
diff --git a/Modules/posixmodule.c b/Modules/posixmodule.c
@@ -4905,42 +4905,34 @@ os_link_impl(PyObject *module, path_t *src, path_t *dst, int src_dir_fd,
 
 
 #if defined(MS_WINDOWS) && !defined(HAVE_OPENDIR)
+static wchar_t *
+join_path_filenameW(const wchar_t *path_wide, const wchar_t *filename,
+                    int normalize);
+
 static PyObject *
 _listdir_windows_no_opendir(path_t *path, PyObject *list)
 {
     PyObject *v;
     HANDLE hFindFile = INVALID_HANDLE_VALUE;
     BOOL result, return_bytes;
-    wchar_t namebuf[MAX_PATH+4]; /* Overallocate for "\*.*" */
-    /* only claim to have space for MAX_PATH */
-    Py_ssize_t len = Py_ARRAY_LENGTH(namebuf)-4;
     wchar_t *wnamebuf = NULL;
 
     WIN32_FIND_DATAW wFileData;
     const wchar_t *po_wchars;
 
     if (!path->wide) { /* Default arg: "." */
         po_wchars = L".";
-        len = 1;
         return_bytes = 0;
     } else {
         po_wchars = path->wide;
-        len = wcslen(path->wide);
         return_bytes = PyBytes_Check(path->object);
     }
-    /* The +5 is so we can append "\\*.*\0" */
-    wnamebuf = PyMem_New(wchar_t, len + 5);
-    if (!wnamebuf) {
-        PyErr_NoMemory();
+
+    wnamebuf = join_path_filenameW(po_wchars, L"*.*", 1);
+    if (wnamebuf == NULL) {
         goto exit;
     }
-    wcscpy(wnamebuf, po_wchars);
-    if (len > 0) {
-        wchar_t wch = wnamebuf[len-1];
-        if (wch != SEP && wch != ALTSEP && wch != L':')
-            wnamebuf[len++] = SEP;
-        wcscpy(wnamebuf + len, L"*.*");
-    }
+
     if ((list = PyList_New(0)) == NULL) {
         goto exit;
     }
@@ -16608,13 +16600,19 @@ static PyType_Spec DirEntryType_spec = {
 
 #ifdef MS_WINDOWS
 
+static int
+is_extended_path(const wchar_t *path)
+{
+    return wcsncmp(path, L"\\\\?\\", 4) == 0;
+}
+
 static wchar_t *
-join_path_filenameW(const wchar_t *path_wide, const wchar_t *filename)
+join_path_filenameW(const wchar_t *path_wide, const wchar_t *filename,
+                    int normalize)
 {
     Py_ssize_t path_len;
-    Py_ssize_t size;
     wchar_t *result;
-    wchar_t ch;
+    wchar_t *path_allocated = NULL;
 
     if (!path_wide) { /* Default arg: "." */
         path_wide = L".";
@@ -16624,20 +16622,44 @@ join_path_filenameW(const wchar_t *path_wide, const wchar_t *filename)
         path_len = wcslen(path_wide);
     }
 
-    /* The +1's are for the path separator and the NUL */
-    size = path_len + 1 + wcslen(filename) + 1;
+    if (path_len == 0) {
+        result = PyMem_New(wchar_t, 1);
+        if (result == NULL) {
+            PyErr_NoMemory();
+            return NULL;
+        }
+        result[0] = L'\0';
+        return result;
+    }
+
+    if (normalize && !is_extended_path(path_wide)) {
+        int err = _PyOS_getfullpathname(path_wide, &path_allocated);
+        if (err < 0) {
+            PyErr_SetFromWindowsErr(0);
+            return NULL;
+        }
+        if (path_allocated == NULL) {
+            PyErr_NoMemory();
+            return NULL;
+        }
+        path_wide = path_allocated;
+        path_len = wcslen(path_wide);
+    }
+
+    size_t size = (size_t)path_len + 1 + wcslen(filename) + 1;
     result = PyMem_New(wchar_t, size);
-    if (!result) {
+    if (result == NULL) {
+        PyMem_RawFree(path_allocated);
         PyErr_NoMemory();
         return NULL;
     }
     wcscpy(result, path_wide);
-    if (path_len > 0) {
-        ch = result[path_len - 1];
-        if (ch != SEP && ch != ALTSEP && ch != L':')
-            result[path_len++] = SEP;
-        wcscpy(result + path_len, filename);
+    wchar_t ch = result[path_len - 1];
+    if (ch != SEP && ch != ALTSEP && ch != L':') {
+        result[path_len++] = SEP;
     }
+    wcscpy(result + path_len, filename);
+    PyMem_RawFree(path_allocated);
     return result;
 }
 
@@ -16669,7 +16691,7 @@ DirEntry_from_find_data(PyObject *module, path_t *path, WIN32_FIND_DATAW *dataW)
             goto error;
     }
 
-    joined_path = join_path_filenameW(path->wide, dataW->cFileName);
+    joined_path = join_path_filenameW(path->wide, dataW->cFileName, 0);
     if (!joined_path)
         goto error;
 
@@ -17105,7 +17127,7 @@ os_scandir_impl(PyObject *module, path_t *path)
 #ifdef MS_WINDOWS
     iterator->first_time = 1;
 
-    path_strW = join_path_filenameW(iterator->path.wide, L"*.*");
+    path_strW = join_path_filenameW(iterator->path.wide, L"*.*", 1);
     if (!path_strW)
         goto error;
 

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+Normalize non-extended Windows paths before appending the wildcard used by`
	`2`	+``os.listdir()`` and ``os.scandir()``, making paths with trailing spaces
	`3`	`+behave consistently with other filesystem APIs.`