gh-92597: Ensure that AST nodes without explicit end positions can be compiled (GH-93359)

(cherry picked from commit 705eaec28f)

Co-authored-by: Pablo Galindo Salgado <Pablogsal@gmail.com>

Lib/test/test_ast.py

@@ -362,6 +362,14 @@ class AST_Tests(unittest.TestCase):
                 with self.assertRaises(ValueError):
                     compile(tree, '<string>', 'exec')
 
+    def test_compilation_of_ast_nodes_with_default_end_position_values(self):
+        tree = ast.Module(body=[
+            ast.Import(names=[ast.alias(name='builtins', lineno=1, col_offset=0)], lineno=1, col_offset=0), 
+            ast.Import(names=[ast.alias(name='traceback', lineno=0, col_offset=0)], lineno=0, col_offset=1)
+        ], type_ignores=[])
+
+        # Check that compilation doesn't crash. Note: this may crash explicitly only on debug mode.
+        compile(tree, "<string>", "exec")
 
     def test_slice(self):
         slc = ast.parse("x[::]").body[0].value.slice

Misc/NEWS.d/next/Core and Builtins/2022-05-30-19-00-38.gh-issue-93359.zXV3A0.rst

@@ -0,0 +1,2 @@
+Ensure that custom :mod:`ast` nodes without explicit end positions can be
+compiled. Patch by Pablo Galindo.

Parser/asdl_c.py

@@ -488,6 +488,12 @@ class Obj2ModPrototypeVisitor(PickleVisitor):
 
 
 class Obj2ModVisitor(PickleVisitor):
+
+    attribute_special_defaults = {
+        "end_lineno": "lineno",
+        "end_col_offset": "col_offset",
+    }
+
     @contextmanager
     def recursive_call(self, node, level):
         self.emit('if (_Py_EnterRecursiveCall(" while traversing \'%s\' node")) {' % node, level, reflow=False)
@@ -637,6 +643,12 @@ class Obj2ModVisitor(PickleVisitor):
             self.emit("if (tmp == NULL || tmp == Py_None) {", depth)
             self.emit("Py_CLEAR(tmp);", depth+1)
             if self.isNumeric(field):
+                if field.name in self.attribute_special_defaults:
+                    self.emit(
+                        "%s = %s;" % (field.name, self.attribute_special_defaults[field.name]),
+                        depth+1,
+                    )
+                else:
                     self.emit("%s = 0;" % field.name, depth+1)
             elif not self.isSimpleType(field):
                 self.emit("%s = NULL;" % field.name, depth+1)

Python/Python-ast.c

@@ -5697,7 +5697,7 @@ obj2ast_stmt(struct ast_state *state, PyObject* obj, stmt_ty* out, PyArena*
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_lineno = 0;
+        end_lineno = lineno;
     }
     else {
         int res;
@@ -5714,7 +5714,7 @@ obj2ast_stmt(struct ast_state *state, PyObject* obj, stmt_ty* out, PyArena*
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_col_offset = 0;
+        end_col_offset = col_offset;
     }
     else {
         int res;
@@ -8114,7 +8114,7 @@ obj2ast_expr(struct ast_state *state, PyObject* obj, expr_ty* out, PyArena*
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_lineno = 0;
+        end_lineno = lineno;
     }
     else {
         int res;
@@ -8131,7 +8131,7 @@ obj2ast_expr(struct ast_state *state, PyObject* obj, expr_ty* out, PyArena*
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_col_offset = 0;
+        end_col_offset = col_offset;
     }
     else {
         int res;
@@ -10291,7 +10291,7 @@ obj2ast_excepthandler(struct ast_state *state, PyObject* obj, excepthandler_ty*
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_lineno = 0;
+        end_lineno = lineno;
     }
     else {
         int res;
@@ -10308,7 +10308,7 @@ obj2ast_excepthandler(struct ast_state *state, PyObject* obj, excepthandler_ty*
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_col_offset = 0;
+        end_col_offset = col_offset;
     }
     else {
         int res;
@@ -10755,7 +10755,7 @@ obj2ast_arg(struct ast_state *state, PyObject* obj, arg_ty* out, PyArena* arena)
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_lineno = 0;
+        end_lineno = lineno;
     }
     else {
         int res;
@@ -10772,7 +10772,7 @@ obj2ast_arg(struct ast_state *state, PyObject* obj, arg_ty* out, PyArena* arena)
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_col_offset = 0;
+        end_col_offset = col_offset;
     }
     else {
         int res;
@@ -10877,7 +10877,7 @@ obj2ast_keyword(struct ast_state *state, PyObject* obj, keyword_ty* out,
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_lineno = 0;
+        end_lineno = lineno;
     }
     else {
         int res;
@@ -10894,7 +10894,7 @@ obj2ast_keyword(struct ast_state *state, PyObject* obj, keyword_ty* out,
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_col_offset = 0;
+        end_col_offset = col_offset;
     }
     else {
         int res;
@@ -10999,7 +10999,7 @@ obj2ast_alias(struct ast_state *state, PyObject* obj, alias_ty* out, PyArena*
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_lineno = 0;
+        end_lineno = lineno;
     }
     else {
         int res;
@@ -11016,7 +11016,7 @@ obj2ast_alias(struct ast_state *state, PyObject* obj, alias_ty* out, PyArena*
     }
     if (tmp == NULL || tmp == Py_None) {
         Py_CLEAR(tmp);
-        end_col_offset = 0;
+        end_col_offset = col_offset;
     }
     else {
         int res;