bpo-40661: Fix segfault when parsing invalid input (GH-20165)

Fix segfaults when parsing very complex invalid input, like `import äˆ ð£„¯ð¢·žð±‹á”€ð””ð‘©±å®ä±¬ð©¾\n𗶽`.

Co-authored-by: Guido van Rossum <guido@python.org>
Co-authored-by: Pablo Galindo <pablogsal@gmail.com>

Lib/test/test_peg_parser.py

@@ -591,6 +591,7 @@ FAIL_TEST_CASES = [
     ("f-string_single_closing_brace", "f'}'"),
     ("from_import_invalid", "from import import a"),
     ("from_import_trailing_comma", "from a import b,"),
+    ("import_non_ascii_syntax_error", "import ä £"),
     # This test case checks error paths involving tokens with uninitialized
     # values of col_offset and end_col_offset.
     ("invalid indentation",

Lib/test/test_syntax.py

@@ -659,6 +659,9 @@ Corner-cases that used to crash:
     Traceback (most recent call last):
     SyntaxError: cannot assign to __debug__
 
+    >>> import ä £
+    Traceback (most recent call last):
+    SyntaxError: invalid character '£' (U+00A3)
 """
 
 import re

Tools/peg_generator/pegen/c_generator.py

@@ -433,6 +433,12 @@ class CParserGenerator(ParserGenerator, GrammarVisitor):
         self.print("int _end_col_offset = _token->end_col_offset;")
         self.print("UNUSED(_end_col_offset); // Only used by EXTRA macro")
 
+    def _check_for_errors(self) -> None:
+        self.print("if (p->error_indicator) {")
+        with self.indent():
+            self.print("return NULL;")
+        self.print("}")
+
     def _set_up_rule_memoization(self, node: Rule, result_type: str) -> None:
         self.print("{")
         with self.indent():
@@ -468,10 +474,7 @@ class CParserGenerator(ParserGenerator, GrammarVisitor):
         memoize = self._should_memoize(node)
 
         with self.indent():
-            self.print("if (p->error_indicator) {")
+            self._check_for_errors()
-            with self.indent():
-                self.print("return NULL;")
-            self.print("}")
             self.print(f"{result_type} _res = NULL;")
             if memoize:
                 self.print(f"if (_PyPegen_is_memoized(p, {node.name}_type, &_res))")
@@ -500,10 +503,7 @@ class CParserGenerator(ParserGenerator, GrammarVisitor):
         is_repeat1 = node.name.startswith("_loop1")
 
         with self.indent():
-            self.print("if (p->error_indicator) {")
+            self._check_for_errors()
-            with self.indent():
-                self.print("return NULL;")
-            self.print("}")
             self.print("void *_res = NULL;")
             if memoize:
                 self.print(f"if (_PyPegen_is_memoized(p, {node.name}_type, &_res))")
@@ -687,6 +687,7 @@ class CParserGenerator(ParserGenerator, GrammarVisitor):
     ) -> None:
         self.print(f"{{ // {node}")
         with self.indent():
+            self._check_for_errors()
             # Prepare variable declarations for the alternative
             vars = self.collect_vars(node)
             for v, var_type in sorted(item for item in vars.items() if item[0] is not None):