[3.11] gh-96670: Raise SyntaxError when parsing NULL bytes (GH-97594) (#104195)

Lib/test/test_ast.py

@@ -857,6 +857,10 @@ class AST_Tests(unittest.TestCase):
         check_limit("a", "[0]")
         check_limit("a", "*a")
 
+    def test_null_bytes(self):
+        with self.assertRaises(SyntaxError,
+            msg="source code string cannot contain null bytes"):
+            ast.parse("a\0b")
 
 class ASTHelpers_Test(unittest.TestCase):
     maxDiff = None

Lib/test/test_builtin.py

@@ -334,11 +334,10 @@ class BuiltinTest(unittest.TestCase):
         self.assertRaises(TypeError, compile)
         self.assertRaises(ValueError, compile, 'print(42)\n', '<string>', 'badmode')
         self.assertRaises(ValueError, compile, 'print(42)\n', '<string>', 'single', 0xff)
-        self.assertRaises(ValueError, compile, chr(0), 'f', 'exec')
         self.assertRaises(TypeError, compile, 'pass', '?', 'exec',
                           mode='eval', source='0', filename='tmp')
         compile('print("\xe5")\n', '', 'exec')
-        self.assertRaises(ValueError, compile, chr(0), 'f', 'exec')
+        self.assertRaises(SyntaxError, compile, chr(0), 'f', 'exec')
         self.assertRaises(ValueError, compile, str('a = 1'), 'f', 'bad')
 
         # test the optimize argument

Lib/test/test_cmd_line_script.py

@@ -657,6 +657,31 @@ class CmdLineTest(unittest.TestCase):
                 ],
             )
 
+    def test_syntaxerror_null_bytes(self):
+        script = "x = '\0' nothing to see here\n';import os;os.system('echo pwnd')\n"
+        with os_helper.temp_dir() as script_dir:
+            script_name = _make_test_script(script_dir, 'script', script)
+            exitcode, stdout, stderr = assert_python_failure(script_name)
+            self.assertEqual(
+                stderr.splitlines()[-2:],
+                [   b"    x = '",
+                    b'SyntaxError: source code cannot contain null bytes'
+                ],
+            )
+
+    def test_syntaxerror_null_bytes_in_multiline_string(self):
+        scripts = ["\n'''\nmultilinestring\0\n'''", "\nf'''\nmultilinestring\0\n'''"] # Both normal and f-strings
+        with os_helper.temp_dir() as script_dir:
+            for script in scripts:
+                script_name = _make_test_script(script_dir, 'script', script)
+                _, _, stderr = assert_python_failure(script_name)
+                self.assertEqual(
+                    stderr.splitlines()[-2:],
+                    [   b"    multilinestring",
+                        b'SyntaxError: source code cannot contain null bytes'
+                    ]
+                )
+
     def test_consistent_sys_path_for_direct_execution(self):
         # This test case ensures that the following all give the same
         # sys.path configuration:

Lib/test/test_compile.py

@@ -542,7 +542,7 @@ if 1:
             with open(fn, "wb") as fp:
                 fp.write(src)
             res = script_helper.run_python_until_end(fn)[0]
-        self.assertIn(b"Non-UTF-8", res.err)
+        self.assertIn(b"source code cannot contain null bytes", res.err)
 
     def test_yet_more_evil_still_undecodable(self):
         # Issue #25388
@@ -552,7 +552,7 @@ if 1:
             with open(fn, "wb") as fp:
                 fp.write(src)
             res = script_helper.run_python_until_end(fn)[0]
-        self.assertIn(b"Non-UTF-8", res.err)
+        self.assertIn(b"source code cannot contain null bytes", res.err)
 
     @support.cpython_only
     def test_compiler_recursion_limit(self):
@@ -588,9 +588,9 @@ if 1:
     def test_null_terminated(self):
         # The source code is null-terminated internally, but bytes-like
         # objects are accepted, which could be not terminated.
-        with self.assertRaisesRegex(ValueError, "cannot contain null"):
+        with self.assertRaisesRegex(SyntaxError, "cannot contain null"):
             compile("123\x00", "<dummy>", "eval")
-        with self.assertRaisesRegex(ValueError, "cannot contain null"):
+        with self.assertRaisesRegex(SyntaxError, "cannot contain null"):
             compile(memoryview(b"123\x00"), "<dummy>", "eval")
         code = compile(memoryview(b"123\x00")[1:-1], "<dummy>", "eval")
         self.assertEqual(eval(code), 23)