gh-139679: Resolve symbolic links to directories when shutil.make_archive() is used to create a zip file

Lib/shutil.py

@@ -1063,10 +1063,28 @@ def _set_uid_gid(tarinfo):
 
 def _make_zipfile(base_name, base_dir, verbose=0, dry_run=0,
                   logger=None, owner=None, group=None, root_dir=None):
-    """Create a zip file from all the files under 'base_dir'.
-
-    The output zip file will be named 'base_name' + ".zip".  Returns the
-    name of the output zip file.
+    """Create a zip file. Returns the name of the zip file.
+
+    - `base_name` is used as the name for the zip file, with a suffix
+      of '.zip'.
+    - `base_dir` is the directory containing the files to be included
+      in the zip file. The zip file is created at the same level in the
+      directory structure as `base_dir`.
+
+    Symbolic links to files and directories are followed and the
+    targets of the links included in the zip file. This matches the
+    default behaviour of command-line zip utilities on
+    Linux/UNIX/Windows systems.
+
+    Hard links to files are followed and the targets of the links
+    included in the zip file. There is no de-duplication of multiple
+    hard links to the same file that is provided by other formats,
+    e.g. tar.
+
+    CAUTION: This function uses os.walk() to prepare the list of files to be
+    included in the zip file. os.walk() does not keep track of which
+    directories have already been visited. A symbolic link to a directory that
+    is a parent of itself will lead to infinite recursion.
     """
     import zipfile  # late import for breaking circular dependency
 
@@ -1094,7 +1112,7 @@ def _make_zipfile(base_name, base_dir, verbose=0, dry_run=0,
                 zf.write(base_dir, arcname)
                 if logger is not None:
                     logger.info("adding '%s'", base_dir)
-            for dirpath, dirnames, filenames in os.walk(base_dir):
+            for dirpath, dirnames, filenames in os.walk(base_dir, followlinks=True):
                 arcdirpath = dirpath
                 if root_dir is not None:
                     arcdirpath = os.path.relpath(arcdirpath, root_dir)
@@ -1183,16 +1201,16 @@ def unregister_archive_format(name):
 
 def make_archive(base_name, format, root_dir=None, base_dir=None, verbose=0,
                  dry_run=0, owner=None, group=None, logger=None):
-    """Create an archive file (eg. zip or tar).
+    """Create an archive file (e.g. zip or tar).
 
     'base_name' is the name of the file to create, minus any format-specific
     extension; 'format' is the archive format: one of "zip", "tar", "gztar",
     "bztar", "xztar", or "zstdtar".  Or any other registered format.
 
     'root_dir' is a directory that will be the root directory of the
-    archive; ie. we typically chdir into 'root_dir' before creating the
+    archive; i.e. we typically chdir into 'root_dir' before creating the
     archive.  'base_dir' is the directory where we start archiving from;
-    ie. 'base_dir' will be the common prefix of all files and
+    i.e. 'base_dir' will be the common prefix of all files and
     directories in the archive.  'root_dir' and 'base_dir' both default
     to the current directory.  Returns the name of the archive file.
 

Lib/test/test_shutil.py

@@ -1636,6 +1636,37 @@ def _create_files(self, base_dir='dist'):
             create_file((root_dir, 'outer'), 'xxx')
         return root_dir, base_dir
 
+    def _create_files_symlinks(self, base_dir="symlinks"):
+        # Create a test structure containing symbolic links to files and
+        # directories.
+        root_dir = self.mkdtemp()
+        wd = os.path.join(root_dir, base_dir)
+        os.mkdir(wd)
+        # Create a regular file.
+        create_file(
+            (wd, 'file1'),
+            'This is file1.'
+            )
+        # Create a symbolic link to the file.
+        os.symlink(
+            os.path.join(wd, 'file1'),
+            os.path.join(wd, 'link1'),
+            )
+        # Create a sub-directory.
+        os.mkdir(os.path.join(wd, 'sub'))
+        # Create a regular file in the sub-directory.
+        create_file(
+            (wd, 'sub', 'file2'),
+            'This is a file2.'
+            )
+        # Create a symbolic link to the sub-directory.
+        os.symlink(
+            os.path.join(wd, 'sub'),
+            os.path.join(wd, 'sub2'),
+            target_is_directory=True,
+            )
+        return root_dir, base_dir
+
     @support.requires_zlib()
     def test_make_tarfile(self):
         root_dir, base_dir = self._create_files()
@@ -1882,6 +1913,47 @@ def test_zipfile_vs_zip(self):
             names2 = zf.namelist()
         self.assertEqual(sorted(names), sorted(names2))
 
+    @os_helper.skip_unless_symlink
+    def test_make_zipfile_symlink_behaviour(self):
+        # Test that symbolic links for both file and directories are resolved
+        # to their targets when shutil.make_archive() is used to make a zip
+        # file, to match default command-line zip behaviour in
+        # Linux/UNIX/Windows.
+        #
+        # If this test is being skipped, it is because either the operating
+        # environment does not support symbolic links, or you do not have the
+        # necessary permissions to create them. For Windows (10 and above) the
+        # test must be invoked from an elevated command prompt or with
+        # Developer Mode turned on.
+        root_dir, base_dir = self._create_files_symlinks()
+        name = os.path.join(root_dir, 'z')
+        archive = shutil.make_archive(
+            name, "zip", root_dir=root_dir, base_dir=base_dir
+            )
+        # Check we have a zip file.
+        self.assertTrue(zipfile.is_zipfile(archive))
+        with zipfile.ZipFile(archive) as zf:
+            extract_dir = os.path.join(root_dir, base_dir, 'extract')
+            zf.extractall(extract_dir)
+        try:
+            # If symbolic link sub2 that targets directory sub1 was
+            # preserved as a link then sub2/file2 will not exist.
+            self.assertTrue(
+                os.path.exists(
+                    os.path.join(extract_dir, base_dir, 'sub2/file2')
+                    )
+                )
+            # If symbolic link link1 that targets file1 was preserved then
+            # it will be a symbolic link.
+            self.assertFalse(
+                os.path.islink(
+                    os.path.join(extract_dir, base_dir, 'link1')
+                    )
+                )
+        finally:
+            # Clean up.
+            shutil.rmtree(root_dir)
+
     @support.requires_zlib()
     @unittest.skipUnless(shutil.which('unzip'),
                          'Need the unzip command to run')

Misc/ACKS

@@ -498,6 +498,7 @@ Eugene Dvurechenski
 Karmen Dykstra
 Josip Dzolonga
 Maxim Dzumanenko
+Thomas Earp
 Hans Eckardt
 Rodolpho Eckhardt
 Ulrich Eckhardt

Misc/NEWS.d/next/Library/2025-10-09-15-50-12.gh-issue-139679.saib2u.rst

@@ -0,0 +1 @@
+The default behaviour of shutil.make_archive(), in relation to handling of symbolic links when creating zip files, now matches the default behaviour of command-line zip utilities on Linux/UNIX/Windows: symbolic links to both files and directories are resolved to their targets and included in the zip file.