Fix typo and broken test cases (#102)

* Fix typo

* Update test cases

* fix(follow Python 3.12.4):

* style: fix black lint complains

data/python.gram

@@ -1805,6 +1805,7 @@ kwargs[list]:
 starred_expression:
  | invalid_starred_expression
  | '*' a=expression { ast.Starred(value=a, ctx=Load, LOCATIONS) }
+ | '*' { self.raise_syntax_error("Invalid star expression") }
 
 kwarg_or_starred:
  | invalid_kwarg
@@ -1921,10 +1922,10 @@ func_type_comment:
 
 # From here on, there are rules for invalid syntax with specialised error messages
 invalid_arguments[NoReturn]:
- | a=args ',' '*' {
- self.raise_syntax_error_known_location(
+ | ((','.(starred_expression | ( assignment_expression | expression !':=') !'=')+ ',' kwargs) | kwargs) a=',' ','.(starred_expression !'=')+ {
+ self.raise_syntax_error_starting_from(
  "iterable argument unpacking follows keyword argument unpacking",
- a[1][-1] if a[1] else a[0][-1],
+ a,
  )
  }
  | a=expression b=for_if_clauses ',' [args | expression for_if_clauses] {

src/pegen/parser_generator.py

@@ -113,7 +113,7 @@ def collect_todo(self) -> None:
  self.todo[rulename].collect_todo(self)
  done = set(alltodo)
 
- def artifical_rule_from_rhs(self, rhs: Rhs) -> str:
+ def artificial_rule_from_rhs(self, rhs: Rhs) -> str:
  self.counter += 1
  name = f"_tmp_{self.counter}" # TODO: Pick a nicer name.
  self.todo[name] = Rule(name, None, rhs)
@@ -129,7 +129,7 @@ def artificial_rule_from_repeat(self, node: Plain, is_repeat1: bool) -> str:
  self.todo[name] = Rule(name, None, Rhs([Alt([NamedItem(None, node)])]))
  return name
 
- def artifical_rule_from_gather(self, node: Gather) -> str:
+ def artificial_rule_from_gather(self, node: Gather) -> str:
  self.counter += 1
  name = f"_gather_{self.counter}"
  self.counter += 1

src/pegen/python_generator.py

@@ -134,7 +134,7 @@ def visit_Rhs(self, node: Rhs) -> Tuple[Optional[str], str]:
  if len(node.alts) == 1 and len(node.alts[0].items) == 1:
  self.cache[node] = self.visit(node.alts[0].items[0])
  else:
- name = self.gen.artifical_rule_from_rhs(node)
+ name = self.gen.artificial_rule_from_rhs(node)
  self.cache[node] = name, f"self.{name}()"
  return self.cache[node]
 
@@ -186,7 +186,7 @@ def visit_Repeat1(self, node: Repeat1) -> Tuple[str, str]:
  def visit_Gather(self, node: Gather) -> Tuple[str, str]:
  if node in self.cache:
  return self.cache[node]
- name = self.gen.artifical_rule_from_gather(node)
+ name = self.gen.artificial_rule_from_gather(node)
  self.cache[node] = name, f"self.{name}()" # No trailing comma here either!
  return self.cache[node]
 

tests/python_parser/conftest.py

@@ -1,4 +1,5 @@
 """"Conftest for pure python parser."""
+
 from pathlib import Path
 
 import pytest

tests/python_parser/test_ast_parsing.py

@@ -1,4 +1,5 @@
 """Test pure Python parser against cpython parser."""
+
 import ast
 import difflib
 import io

tests/python_parser/test_syntax_error_handling.py

@@ -1,4 +1,5 @@
 """Test syntax errors for cases where the parser can generate helpful messages."""
+
 import sys
 
 import pytest
@@ -93,57 +94,71 @@ def parse_invalid_syntax(
  [
  (
  "f'a = {}'",
- "valid expression required before '}'"
- if sys.version_info >= (3, 12)
- else "f-string: empty expression not allowed",
+ (
+ "valid expression required before '}'"
+ if sys.version_info >= (3, 12)
+ else "f-string: empty expression not allowed"
+ ),
  (1, 8) if sys.version_info >= (3, 12) else None,
  (1, 9) if sys.version_info >= (3, 12) else None,
  ),
  (
  "f'a = {=}'",
- "expression required before '='"
- if sys.version_info >= (3, 11)
- else "f-string: empty expression not allowed",
+ (
+ "expression required before '='"
+ if sys.version_info >= (3, 11)
+ else "f-string: empty expression not allowed"
+ ),
  (1, 8) if sys.version_info >= (3, 12) else None,
  (1, 9) if sys.version_info >= (3, 12) else None,
  ),
  (
  "f'a = {!}'",
- "expression required before '!'"
- if sys.version_info >= (3, 11)
- else "f-string: empty expression not allowed",
+ (
+ "expression required before '!'"
+ if sys.version_info >= (3, 11)
+ else "f-string: empty expression not allowed"
+ ),
  (1, 8) if sys.version_info >= (3, 12) else None,
  (1, 9) if sys.version_info >= (3, 12) else None,
  ),
  (
  "f'a = {:}'",
- "expression required before ':'"
- if sys.version_info >= (3, 11)
- else "f-string: empty expression not allowed",
+ (
+ "expression required before ':'"
+ if sys.version_info >= (3, 11)
+ else "f-string: empty expression not allowed"
+ ),
  (1, 8) if sys.version_info >= (3, 12) else None,
  (1, 9) if sys.version_info >= (3, 12) else (1, 11),
  ),
  (
  "f'a = {a=d}'",
- "expecting '!', or ':', or '}'"
- if sys.version_info >= (3, 12)
- else "f-string: expecting '}'",
+ (
+ "expecting '!', or ':', or '}'"
+ if sys.version_info >= (3, 12)
+ else "f-string: expecting '}'"
+ ),
  (1, 10) if sys.version_info >= (3, 12) else None,
  (1, 11) if sys.version_info >= (3, 12) else None,
  ),
  (
  "f'a = { 1 + }'",
- "f-string: expecting '=', or '!', or ':', or '}'"
- if sys.version_info >= (3, 12)
- else "f-string: invalid syntax",
+ (
+ "f-string: expecting '=', or '!', or ':', or '}'"
+ if sys.version_info >= (3, 12)
+ else "f-string: invalid syntax"
+ ),
  (1, 11) if sys.version_info >= (3, 12) else (1, 7),
  (1, 12) if sys.version_info >= (3, 12) else (1, 8),
  ),
  (
  "(\n\t'b'\n\tf'a = { 1 + }'\n)",
- "f-string: expecting '=', or '!', or ':', or '}'"
- if sys.version_info >= (3, 12)
- else "f-string: invalid syntax",
+ (
+ "f-string: expecting '=', or '!', or ':', or '}'"
+ if sys.version_info >= (3, 12)
+ else "f-string: invalid syntax"
+ ),
  (3, 12) if sys.version_info >= (3, 12) else (3, 7),
  (3, 13) if sys.version_info >= (3, 12) else (3, 8),
  ),
@@ -210,8 +225,8 @@ def test_invalid_statements(
  (
  "f(**a, *b)",
  "iterable argument unpacking follows keyword argument unpacking",
- (1, 3),
- (1, 6),
+ (1, 6) if sys.version_info >= (3, 12) else None,
+ (1, 10) if sys.version_info >= (3, 12) else None,
  ),
  # NOTE CPython bug, should report 15 as expected (we use None to omit the check)
  ("f(a for a in b, c)", "Generator expression must be parenthesized", (1, 3), (1, None)),

tests/python_parser/test_unsupported_syntax.py

@@ -4,6 +4,7 @@
 not broader since we would not be able to generate the proper ast nodes.
 
 """
+
 import io
 import tokenize