gh-105858: Improve AST node constructors (#105880)

Demonstration:

>>> ast.FunctionDef.__annotations__
{'name': <class 'str'>, 'args': <class 'ast.arguments'>, 'body': list[ast.stmt], 'decorator_list': list[ast.expr], 'returns': ast.expr | None, 'type_comment': str | None, 'type_params': list[ast.type_param]}
>>> ast.FunctionDef()
<stdin>:1: DeprecationWarning: FunctionDef.__init__ missing 1 required positional argument: 'name'. This will become an error in Python 3.15.
<stdin>:1: DeprecationWarning: FunctionDef.__init__ missing 1 required positional argument: 'args'. This will become an error in Python 3.15.
<ast.FunctionDef object at 0x101959460>
>>> node = ast.FunctionDef(name="foo", args=ast.arguments())
>>> node.decorator_list
[]
>>> ast.FunctionDef(whatever="you want", name="x", args=ast.arguments())
<stdin>:1: DeprecationWarning: FunctionDef.__init__ got an unexpected keyword argument 'whatever'. Support for arbitrary keyword arguments is deprecated and will be removed in Python 3.15.
<ast.FunctionDef object at 0x1019581f0>

Author: JelleZijlstra

Doc/library/ast.rst

@@ -103,20 +103,15 @@ Node classes
    For example, to create and populate an :class:`ast.UnaryOp` node, you could
    use ::
 
-      node = ast.UnaryOp()
-      node.op = ast.USub()
-      node.operand = ast.Constant()
-      node.operand.value = 5
-      node.operand.lineno = 0
-      node.operand.col_offset = 0
-      node.lineno = 0
-      node.col_offset = 0
-
-   or the more compact ::
-
       node = ast.UnaryOp(ast.USub(), ast.Constant(5, lineno=0, col_offset=0),
                          lineno=0, col_offset=0)
 
+   If a field that is optional in the grammar is omitted from the constructor,
+   it defaults to ``None``. If a list field is omitted, it defaults to the empty
+   list. If any other field is omitted, a :exc:`DeprecationWarning` is raised
+   and the AST node will not have this field. In Python 3.15, this condition will
+   raise an error.
+
 .. versionchanged:: 3.8
 
    Class :class:`ast.Constant` is now used for all constants.
@@ -140,6 +135,14 @@ Node classes
    In the meantime, instantiating them will return an instance of
    a different class.
 
+.. deprecated-removed:: 3.13 3.15
+
+   Previous versions of Python allowed the creation of AST nodes that were missing
+   required fields. Similarly, AST node constructors allowed arbitrary keyword
+   arguments that were set as attributes of the AST node, even if they did not
+   match any of the fields of the AST node. This behavior is deprecated and will
+   be removed in Python 3.15.
+
 .. note::
     The descriptions of the specific node classes displayed here
     were initially adapted from the fantastic `Green Tree

Doc/whatsnew/3.13.rst

@@ -206,6 +206,21 @@ array
 ast
 ---
 
+* The constructors of node types in the :mod:`ast` module are now stricter
+  in the arguments they accept, and have more intuitive behaviour when
+  arguments are omitted.
+
+  If an optional field on an AST node is not included as an argument when
+  constructing an instance, the field will now be set to ``None``. Similarly,
+  if a list field is omitted, that field will now be set to an empty list.
+  (Previously, in both cases, the attribute would be missing on the newly
+  constructed AST node instance.)
+
+  If other arguments are omitted, a :exc:`DeprecationWarning` is emitted.
+  This will cause an exception in Python 3.15. Similarly, passing a keyword
+  argument that does not map to a field on the AST node is now deprecated,
+  and will raise an exception in Python 3.15.
+
 * :func:`ast.parse` now accepts an optional argument ``optimize``
   which is passed on to the :func:`compile` built-in. This makes it
   possible to obtain an optimized ``AST``.

Include/internal/pycore_global_objects_fini_generated.h

@@ -242,6 +242,7 @@ struct _Py_global_strings {
         STRUCT_FOR_ID(_check_retval_)
         STRUCT_FOR_ID(_dealloc_warn)
         STRUCT_FOR_ID(_feature_version)
+        STRUCT_FOR_ID(_field_types)
         STRUCT_FOR_ID(_fields_)
         STRUCT_FOR_ID(_finalizing)
         STRUCT_FOR_ID(_find_and_load)

Include/internal/pycore_global_strings.h

@@ -525,17 +525,38 @@ def test_field_attr_existence(self):
             if name == 'Index':
                 continue
             if self._is_ast_node(name, item):
-                x = item()
+                x = self._construct_ast_class(item)
                 if isinstance(x, ast.AST):
                     self.assertIs(type(x._fields), tuple)
 
+    def _construct_ast_class(self, cls):
+        kwargs = {}
+        for name, typ in cls.__annotations__.items():
+            if typ is str:
+                kwargs[name] = 'capybara'
+            elif typ is int:
+                kwargs[name] = 42
+            elif typ is object:
+                kwargs[name] = b'capybara'
+            elif isinstance(typ, type) and issubclass(typ, ast.AST):
+                kwargs[name] = self._construct_ast_class(typ)
+        return cls(**kwargs)
+
     def test_arguments(self):
         x = ast.arguments()
         self.assertEqual(x._fields, ('posonlyargs', 'args', 'vararg', 'kwonlyargs',
                                      'kw_defaults', 'kwarg', 'defaults'))
-
-        with self.assertRaises(AttributeError):
-            x.args
+        self.assertEqual(x.__annotations__, {
+            'posonlyargs': list[ast.arg],
+            'args': list[ast.arg],
+            'vararg': ast.arg | None,
+            'kwonlyargs': list[ast.arg],
+            'kw_defaults': list[ast.expr],
+            'kwarg': ast.arg | None,
+            'defaults': list[ast.expr],
+        })
+
+        self.assertEqual(x.args, [])
         self.assertIsNone(x.vararg)
 
         x = ast.arguments(*range(1, 8))
@@ -551,7 +572,7 @@ def test_field_attr_writable_deprecated(self):
         self.assertEqual(x._fields, 666)
 
     def test_field_attr_writable(self):
-        x = ast.Constant()
+        x = ast.Constant(1)
         # We can assign to _fields
         x._fields = 666
         self.assertEqual(x._fields, 666)
@@ -611,15 +632,22 @@ def test_classattrs_deprecated(self):
 
         self.assertEqual([str(w.message) for w in wlog], [
             'ast.Num is deprecated and will be removed in Python 3.14; use ast.Constant instead',
+            "Constant.__init__ missing 1 required positional argument: 'value'. This will become "
+            'an error in Python 3.15.',
             'Attribute n is deprecated and will be removed in Python 3.14; use value instead',
             'ast.Num is deprecated and will be removed in Python 3.14; use ast.Constant instead',
             'Attribute n is deprecated and will be removed in Python 3.14; use value instead',
             'ast.Num is deprecated and will be removed in Python 3.14; use ast.Constant instead',
+            "Constant.__init__ missing 1 required positional argument: 'value'. This will become "
+            'an error in Python 3.15.',
             'ast.Num is deprecated and will be removed in Python 3.14; use ast.Constant instead',
             'Attribute n is deprecated and will be removed in Python 3.14; use value instead',
             'ast.Num is deprecated and will be removed in Python 3.14; use ast.Constant instead',
             'ast.Num is deprecated and will be removed in Python 3.14; use ast.Constant instead',
             'ast.Num is deprecated and will be removed in Python 3.14; use ast.Constant instead',
+            "Constant.__init__ got an unexpected keyword argument 'foo'. Support for "
+            'arbitrary keyword arguments is deprecated and will be removed in Python '
+            '3.15.',
             'ast.Num is deprecated and will be removed in Python 3.14; use ast.Constant instead',
             'Attribute n is deprecated and will be removed in Python 3.14; use value instead',
             'ast.Num is deprecated and will be removed in Python 3.14; use ast.Constant instead',
@@ -636,7 +664,8 @@ def test_classattrs_deprecated(self):
         ])
 
     def test_classattrs(self):
-        x = ast.Constant()
+        with self.assertWarns(DeprecationWarning):
+            x = ast.Constant()
         self.assertEqual(x._fields, ('value', 'kind'))
 
         with self.assertRaises(AttributeError):
@@ -651,7 +680,7 @@ def test_classattrs(self):
         with self.assertRaises(AttributeError):
             x.foobar
 
-        x = ast.Constant(lineno=2)
+        x = ast.Constant(lineno=2, value=3)
         self.assertEqual(x.lineno, 2)
 
         x = ast.Constant(42, lineno=0)
@@ -662,8 +691,9 @@ def test_classattrs(self):
         self.assertRaises(TypeError, ast.Constant, 1, None, 2)
         self.assertRaises(TypeError, ast.Constant, 1, None, 2, lineno=0)
 
-        # Arbitrary keyword arguments are supported
-        self.assertEqual(ast.Constant(1, foo='bar').foo, 'bar')
+        # Arbitrary keyword arguments are supported (but deprecated)
+        with self.assertWarns(DeprecationWarning):
+            self.assertEqual(ast.Constant(1, foo='bar').foo, 'bar')
 
         with self.assertRaisesRegex(TypeError, "Constant got multiple values for argument 'value'"):
             ast.Constant(1, value=2)
@@ -815,11 +845,11 @@ def test_isinstance(self):
         assertBytesDeprecated(self.assertNotIsInstance, Constant('42'), Bytes)
         assertNameConstantDeprecated(self.assertNotIsInstance, Constant(42), NameConstant)
         assertEllipsisDeprecated(self.assertNotIsInstance, Constant(42), Ellipsis)
-        assertNumDeprecated(self.assertNotIsInstance, Constant(), Num)
-        assertStrDeprecated(self.assertNotIsInstance, Constant(), Str)
-        assertBytesDeprecated(self.assertNotIsInstance, Constant(), Bytes)
-        assertNameConstantDeprecated(self.assertNotIsInstance, Constant(), NameConstant)
-        assertEllipsisDeprecated(self.assertNotIsInstance, Constant(), Ellipsis)
+        assertNumDeprecated(self.assertNotIsInstance, Constant(None), Num)
+        assertStrDeprecated(self.assertNotIsInstance, Constant(None), Str)
+        assertBytesDeprecated(self.assertNotIsInstance, Constant(None), Bytes)
+        assertNameConstantDeprecated(self.assertNotIsInstance, Constant(1), NameConstant)
+        assertEllipsisDeprecated(self.assertNotIsInstance, Constant(None), Ellipsis)
 
         class S(str): pass
         with assertStrDeprecated():
@@ -888,8 +918,9 @@ def test_module(self):
         self.assertEqual(x.body, body)
 
     def test_nodeclasses(self):
-        # Zero arguments constructor explicitly allowed
-        x = ast.BinOp()
+        # Zero arguments constructor explicitly allowed (but deprecated)
+        with self.assertWarns(DeprecationWarning):
+            x = ast.BinOp()
         self.assertEqual(x._fields, ('left', 'op', 'right'))
 
         # Random attribute allowed too
@@ -927,8 +958,9 @@ def test_nodeclasses(self):
         self.assertEqual(x.right, 3)
         self.assertEqual(x.lineno, 0)
 
-        # Random kwargs also allowed
-        x = ast.BinOp(1, 2, 3, foobarbaz=42)
+        # Random kwargs also allowed (but deprecated)
+        with self.assertWarns(DeprecationWarning):
+            x = ast.BinOp(1, 2, 3, foobarbaz=42)
         self.assertEqual(x.foobarbaz, 42)
 
     def test_no_fields(self):
@@ -941,8 +973,9 @@ def test_pickling(self):
 
         for protocol in range(pickle.HIGHEST_PROTOCOL + 1):
             for ast in (compile(i, "?", "exec", 0x400) for i in exec_tests):
-                ast2 = pickle.loads(pickle.dumps(ast, protocol))
-                self.assertEqual(to_tuple(ast2), to_tuple(ast))
+                with self.subTest(ast=ast, protocol=protocol):
+                    ast2 = pickle.loads(pickle.dumps(ast, protocol))
+                    self.assertEqual(to_tuple(ast2), to_tuple(ast))
 
     def test_invalid_sum(self):
         pos = dict(lineno=2, col_offset=3)
@@ -1310,8 +1343,9 @@ def test_copy_location(self):
             'lineno=1, col_offset=4, end_lineno=1, end_col_offset=5), lineno=1, '
             'col_offset=0, end_lineno=1, end_col_offset=5))'
         )
-        src = ast.Call(col_offset=1, lineno=1, end_lineno=1, end_col_offset=1)
-        new = ast.copy_location(src, ast.Call(col_offset=None, lineno=None))
+        func = ast.Name('spam', ast.Load())
+        src = ast.Call(col_offset=1, lineno=1, end_lineno=1, end_col_offset=1, func=func)
+        new = ast.copy_location(src, ast.Call(col_offset=None, lineno=None, func=func))
         self.assertIsNone(new.end_lineno)
         self.assertIsNone(new.end_col_offset)
         self.assertEqual(new.lineno, 1)
@@ -1570,15 +1604,15 @@ def test_level_as_none(self):
         self.assertIn('sleep', ns)
 
     def test_recursion_direct(self):
-        e = ast.UnaryOp(op=ast.Not(), lineno=0, col_offset=0)
+        e = ast.UnaryOp(op=ast.Not(), lineno=0, col_offset=0, operand=ast.Constant(1))
         e.operand = e
         with self.assertRaises(RecursionError):
             with support.infinite_recursion():
                 compile(ast.Expression(e), "<test>", "eval")
 
     def test_recursion_indirect(self):
-        e = ast.UnaryOp(op=ast.Not(), lineno=0, col_offset=0)
-        f = ast.UnaryOp(op=ast.Not(), lineno=0, col_offset=0)
+        e = ast.UnaryOp(op=ast.Not(), lineno=0, col_offset=0, operand=ast.Constant(1))
+        f = ast.UnaryOp(op=ast.Not(), lineno=0, col_offset=0, operand=ast.Constant(1))
         e.operand = f
         f.operand = e
         with self.assertRaises(RecursionError):
@@ -2866,6 +2900,23 @@ def visit_Call(self, node: ast.Call):
         self.assertASTTransformation(PrintToLog, code, expected)
 
 
+class ASTConstructorTests(unittest.TestCase):
+    """Test the autogenerated constructors for AST nodes."""
+
+    def test_FunctionDef(self):
+        args = ast.arguments()
+        self.assertEqual(args.args, [])
+        self.assertEqual(args.posonlyargs, [])
+        with self.assertWarnsRegex(DeprecationWarning,
+                                   r"FunctionDef\.__init__ missing 1 required positional argument: 'name'"):
+            node = ast.FunctionDef(args=args)
+        self.assertFalse(hasattr(node, "name"))
+        self.assertEqual(node.decorator_list, [])
+        node = ast.FunctionDef(name='foo', args=args)
+        self.assertEqual(node.name, 'foo')
+        self.assertEqual(node.decorator_list, [])
+
+
 @support.cpython_only
 class ModuleStateTests(unittest.TestCase):
     # bpo-41194, bpo-41261, bpo-41631: The _ast module uses a global state.

Include/internal/pycore_runtime_init_generated.h

@@ -0,0 +1,8 @@
+Improve the constructors for :mod:`ast` nodes. Arguments of list types now
+default to an empty list if omitted, and optional fields default to ``None``.
+AST nodes now have an
+``__annotations__`` attribute with the expected types of their attributes.
+Passing unrecognized extra arguments to AST nodes is deprecated and will
+become an error in Python 3.15. Omitting a required argument to an AST node
+is deprecated and will become an error in Python 3.15. Patch by Jelle
+Zijlstra.