Syntax error suggestions

The behavior seems good but not sure about the stuff codex did to
the doctests

Author: JelleZijlstra

Lib/test/test_syntax.py

@@ -353,10 +353,14 @@
 # Make sure soft keywords constructs don't raise specialized
 # errors regarding missing commas or other spezialiced errors
 
->>> match x:
-...     y = 3
-Traceback (most recent call last):
-SyntaxError: invalid syntax
+>>> import linecache, traceback
+
+>>> try:
+...     match x:
+...         y = 3
+... except NameError as e:
+...     print(traceback.format_exception(e)[-1], end="")
+NameError: name 'match' is not defined. Did you mean to use a 'match' statement with 'case' clauses?
 
 >>> match x:
 ...     case y:
@@ -394,13 +398,17 @@
 Traceback (most recent call last):
 SyntaxError: case statement must be inside match statement
 
->>> case match: ...
-Traceback (most recent call last):
-SyntaxError: case statement must be inside match statement
+>>> try:
+...     case match: ...
+... except NameError as e:
+...     print(traceback.format_exception(e)[-1], end="")
+NameError: name 'case' is not defined. Did you mean to use a 'case' pattern inside a 'match' statement?
 
->>> case case: ...
-Traceback (most recent call last):
-SyntaxError: case statement must be inside match statement
+>>> try:
+...     case case: ...
+... except NameError as e:
+...     print(traceback.format_exception(e)[-1], end="")
+NameError: name 'case' is not defined. Did you mean to use a 'case' pattern inside a 'match' statement?
 
 >>> if some:
 ...     case 1: ...
@@ -1832,12 +1840,14 @@
 Traceback (most recent call last):
 SyntaxError: invalid syntax. Did you mean 'if'?
 
->>> if x:
-...   pass
-... elseif y:
-...   pass
-Traceback (most recent call last):
-SyntaxError: invalid syntax. Did you mean 'elif'?
+>>> try:
+...   source = "x = y = 1\\nif x:\\n  pass\\nelseif y:\\n  pass"
+...   filename = "<elseif-builder-test>"
+...   linecache.cache[filename] = (len(source), None, source.splitlines(True), filename)
+...   exec(compile(source, filename, "exec"))
+... except NameError as e:
+...   print(traceback.format_exception(e)[-1], end="")
+NameError: name 'elseif' is not defined. Did you mean: 'elif'?
 
 >>> if x:
 ...   pass
@@ -1855,10 +1865,12 @@
 Traceback (most recent call last):
 SyntaxError: invalid syntax. Did you mean 'try'?
 
->>> classe MyClass:
-...   pass
-Traceback (most recent call last):
-SyntaxError: invalid syntax. Did you mean 'class'?
+>>> try:
+...   classe MyClass:
+...     pass
+... except NameError as e:
+...   print(traceback.format_exception(e)[-1], end="")
+NameError: name 'classe' is not defined. Did you mean: 'class'?
 
 >>> impor math
 Traceback (most recent call last):
@@ -1871,16 +1883,18 @@
 >>> defn calculate_sum(a, b):
 ...   return a + b
 Traceback (most recent call last):
-SyntaxError: invalid syntax. Did you mean 'def'?
+SyntaxError: 'return' outside function
 
 >>> def foo():
 ...   returm result
 Traceback (most recent call last):
 SyntaxError: invalid syntax. Did you mean 'return'?
 
->>> lamda x: x ** 2
-Traceback (most recent call last):
-SyntaxError: invalid syntax. Did you mean 'lambda'?
+>>> try:
+...   lamda x: x ** 2
+... except NameError as e:
+...   print(traceback.format_exception(e)[-1], end="")
+NameError: name 'lamda' is not defined. Did you mean: 'lambda'?
 
 >>> def foo():
 ...   yeld i

Lib/test/test_traceback.py

@@ -1798,14 +1798,10 @@ class TestKeywordTypoSuggestions(unittest.TestCase):
         ("for a in b:\n  pass\nelso:\n  pass", "else"),
         ("whille True:\n  pass", "while"),
         ("iff x > 5:\n  pass", "if"),
-        ("if x:\n  pass\nelseif y:\n  pass", "elif"),
         ("tyo:\n  pass\nexcept y:\n  pass", "try"),
-        ("classe MyClass:\n  pass", "class"),
         ("impor math", "import"),
         ("form x import y", "from"),
-        ("defn calculate_sum(a, b):\n  return a + b", "def"),
         ("def foo():\n  returm result", "return"),
-        ("lamda x: x ** 2", "lambda"),
         ("def foo():\n  yeld i", "yield"),
         ("def foo():\n  globel counter", "global"),
         ("frum math import sqrt", "from"),
@@ -1816,6 +1812,10 @@ class TestKeywordTypoSuggestions(unittest.TestCase):
         ("[123 fur x\nin range(3)\nif x]", "for"),
         ("for x im n:\n  pass", "in"),
     ]
+    RUNTIME_TYPO_CASES = [
+        ("classe MyClass:\n  pass", "Did you mean: 'class'"),
+        ("lamda x: x ** 2", "Did you mean: 'lambda'"),
+    ]
 
     def test_keyword_suggestions_from_file(self):
         with tempfile.TemporaryDirectory() as script_dir:
@@ -1826,6 +1826,13 @@ def test_keyword_suggestions_from_file(self):
                     rc, stdout, stderr = assert_python_failure(script_name)
                     stderr_text = stderr.decode('utf-8')
                     self.assertIn(f"Did you mean '{expected_kw}'", stderr_text)
+            for i, (code, expected) in enumerate(self.RUNTIME_TYPO_CASES):
+                with self.subTest(typo=expected):
+                    source = textwrap.dedent(code).strip()
+                    script_name = make_script(script_dir, f"runtime_script_{i}", source)
+                    rc, stdout, stderr = assert_python_failure(script_name)
+                    stderr_text = stderr.decode('utf-8')
+                    self.assertIn(expected, stderr_text)
 
     def test_keyword_suggestions_from_command_string(self):
         for code, expected_kw in self.TYPO_CASES:
@@ -1834,6 +1841,12 @@ def test_keyword_suggestions_from_command_string(self):
                 rc, stdout, stderr = assert_python_failure('-c', source)
                 stderr_text = stderr.decode('utf-8')
                 self.assertIn(f"Did you mean '{expected_kw}'", stderr_text)
+        for code, expected in self.RUNTIME_TYPO_CASES:
+            with self.subTest(typo=expected):
+                source = textwrap.dedent(code).strip()
+                rc, stdout, stderr = assert_python_failure('-c', source)
+                stderr_text = stderr.decode('utf-8')
+                self.assertIn(expected, stderr_text)
 
     def test_no_keyword_suggestion_for_comma_errors(self):
         # When the parser identifies a missing comma, don't suggest
@@ -5094,6 +5107,32 @@ def func():
         actual = self.get_suggestion(func)
         self.assertIn("forget to import '_io'", actual)
 
+    def test_name_error_for_class_builder_keyword_typos(self):
+        def func():
+            source = "classe MyClass:\n    pass"
+            filename = "<class-builder-keyword-typo-test>"
+            linecache.cache[filename] = (
+                len(source), None, source.splitlines(True), filename)
+            exec(compile(source, filename, "exec"))
+
+        actual = self.get_suggestion(func)
+        self.assertIn("Did you mean: 'class'?", actual)
+
+    def test_name_error_for_class_builder_soft_keywords(self):
+        def func():
+            source = "case match: ..."
+            filename = "<class-builder-soft-keyword-test>"
+            linecache.cache[filename] = (
+                len(source), None, source.splitlines(True), filename)
+            exec(compile(source, filename, "exec"))
+
+        actual = self.get_suggestion(func)
+        self.assertIn(
+            "Did you mean to use a 'case' pattern inside a 'match' statement?",
+            actual,
+        )
+        self.assertNotIn("'False'", actual)
+
 
 
 class PurePythonSuggestionFormattingTests(

Lib/traceback.py

@@ -1234,7 +1234,14 @@ def __init__(self, exc_type, exc_value, exc_traceback, *, limit=None,
         elif exc_type and issubclass(exc_type, NameError) and \
                 getattr(exc_value, "name", None) is not None:
             wrong_name = getattr(exc_value, "name", None)
-            suggestion = _compute_suggestion_error(exc_value, exc_traceback, wrong_name)
+            builder_hint = _compute_class_builder_name_error_hint(
+                wrong_name, self.stack)
+            suggestion = None
+            if builder_hint:
+                self._str += f". {builder_hint}"
+            else:
+                suggestion = _compute_suggestion_error(
+                    exc_value, exc_traceback, wrong_name)
             if suggestion:
                 if suggestion.isascii():
                     self._str += f". Did you mean: '{suggestion}'?"
@@ -1994,6 +2001,63 @@ def _compute_suggestion_error(exc_value, tb, wrong_name):
     return suggestion
 
 
+def _tokenize_line(line):
+    try:
+        return list(tokenize.generate_tokens(io.StringIO(line).readline))
+    except tokenize.TokenError:
+        return []
+
+
+def _looks_like_class_builder_header(line, wrong_name):
+    tokens = [
+        token for token in _tokenize_line(line)
+        if token.type not in {
+            tokenize.ENCODING, tokenize.INDENT, tokenize.DEDENT,
+            tokenize.NL, tokenize.NEWLINE, tokenize.ENDMARKER,
+        }
+    ]
+    if len(tokens) < 3:
+        return False
+    if tokens[0].type != tokenize.NAME or tokens[0].string != wrong_name:
+        return False
+    if tokens[1].type != tokenize.NAME:
+        return False
+    return tokens[2].string in {":", "(", "["}
+
+
+def _compute_class_builder_name_error_hint(wrong_name, stack):
+    if wrong_name is None or not isinstance(wrong_name, str):
+        return None
+    if not stack:
+        return None
+    line = stack[-1].line
+    if not line or not _looks_like_class_builder_header(line, wrong_name):
+        return None
+
+    if wrong_name == "match":
+        return "Did you mean to use a 'match' statement with 'case' clauses?"
+    if wrong_name == "case":
+        return "Did you mean to use a 'case' pattern inside a 'match' statement?"
+
+    candidates = [name for name in keyword.kwlist + keyword.softkwlist
+                  if name != "_"]
+    try:
+        import _suggestions
+    except ImportError:
+        pass
+    else:
+        suggestion = _suggestions._generate_suggestions(candidates, wrong_name)
+        if suggestion:
+            return f"Did you mean: '{suggestion}'?"
+
+    import difflib
+    suggestions = difflib.get_close_matches(
+        wrong_name, candidates, n=1, cutoff=0.5)
+    if suggestions:
+        return f"Did you mean: '{suggestions[0]}'?"
+    return None
+
+
 def _levenshtein_distance(a, b, max_cost):
     # A Python implementation of Python/suggestions.c:levenshtein_distance.