gh-89727: Fix pathlib.Path.walk RecursionError on deep trees

Lib/pathlib.py

@@ -19,6 +19,7 @@
 from operator import attrgetter
 from stat import S_ISDIR, S_ISLNK, S_ISREG, S_ISSOCK, S_ISBLK, S_ISCHR, S_ISFIFO
 from urllib.parse import quote_from_bytes as urlquote_from_bytes
+from collections import deque
 
 
 __all__ = [
@@ -1279,45 +1280,49 @@ def expanduser(self):
     def walk(self, top_down=True, on_error=None, follow_symlinks=False):
         """Walk the directory tree from this directory, similar to os.walk()."""
         sys.audit("pathlib.Path.walk", self, on_error, follow_symlinks)
-        return self._walk(top_down, on_error, follow_symlinks)
-
-    def _walk(self, top_down, on_error, follow_symlinks):
-        # We may not have read permission for self, in which case we can't
-        # get a list of the files the directory contains. os.walk
-        # always suppressed the exception then, rather than blow up for a
-        # minor reason when (say) a thousand readable directories are still
-        # left to visit. That logic is copied here.
-        try:
-            scandir_it = self._scandir()
-        except OSError as error:
-            if on_error is not None:
-                on_error(error)
-            return
-
-        with scandir_it:
-            dirnames = []
-            filenames = []
-            for entry in scandir_it:
-                try:
-                    is_dir = entry.is_dir(follow_symlinks=follow_symlinks)
-                except OSError:
-                    # Carried over from os.path.isdir().
-                    is_dir = False
-
-                if is_dir:
-                    dirnames.append(entry.name)
-                else:
-                    filenames.append(entry.name)
-
-        if top_down:
-            yield self, dirnames, filenames
-
-        for dirname in dirnames:
-            dirpath = self._make_child_relpath(dirname)
-            yield from dirpath._walk(top_down, on_error, follow_symlinks)
+        stack = [(False, self)]
+
+        while stack:
+            must_yield, top = stack.pop()
+            if must_yield:
+                yield top
+                continue
+
+            # We may not have read permission for self, in which case we can't
+            # get a list of the files the directory contains. os.walk
+            # always suppressed the exception then, rather than blow up for a
+            # minor reason when (say) a thousand readable directories are still
+            # left to visit. That logic is copied here.
+            try:
+                scandir_it = top._scandir()
+            except OSError as error:
+                if on_error is not None:
+                    on_error(error)
+                continue
+
+            with scandir_it:
+                dirnames = []
+                filenames = []
+                for entry in scandir_it:
+                    try:
+                        is_dir = entry.is_dir(follow_symlinks=follow_symlinks)
+                    except OSError:
+                        # Carried over from os.path.isdir().
+                        is_dir = False
+
+                    if is_dir:
+                        dirnames.append(entry.name)
+                    else:
+                        filenames.append(entry.name)
+
+            if top_down:
+                yield top, dirnames, filenames
+            else:
+                stack.append((True, (top, dirnames, filenames)))
 
-        if not top_down:
-            yield self, dirnames, filenames
+            for dirname in reversed(dirnames):
+                dirpath = top._make_child_relpath(dirname)
+                stack.append((False, dirpath))
 
 
 class PosixPath(Path, PurePosixPath):

Lib/test/test_pathlib.py

@@ -13,6 +13,7 @@
 from unittest import mock
 
 from test.support import import_helper
+from test.support import infinite_recursion
 from test.support import is_emscripten, is_wasi
 from test.support import os_helper
 from test.support.os_helper import TESTFN, FakePath
@@ -2766,6 +2767,15 @@ def test_walk_many_open_files(self):
                 self.assertEqual(next(it), expected)
             path = path / 'd'
 
+    def test_walk_above_recursion_limit(self):
+        base = pathlib.Path(os_helper.TESTFN, 'deep')
+        path = pathlib.Path(base, *(['d']*50))
+        path.mkdir(parents=True)
+
+        with infinite_recursion(40):
+            list(base.walk())
+            list(base.walk(top_down=False))
+
 
 class PathTest(_BasePathTest, unittest.TestCase):
     cls = pathlib.Path

Misc/NEWS.d/next/Library/2022-12-16-10-27-58.gh-issue-89727.y64ZLM.rst

@@ -0,0 +1 @@
+Fix pathlib.Path.walk RecursionError on deep directory trees by rewriting it using iteration instead of recursion.