gh-104504: Run mypy on cases_generator in CI (and blacken the code) (gh-108090)

Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>

.github/dependabot.yml

@@ -13,7 +13,7 @@ updates:
           - "version-update:semver-minor"
           - "version-update:semver-patch"
   - package-ecosystem: "pip"
-    directory: "/Tools/clinic/"
+    directory: "/Tools/"
     schedule:
       interval: "monthly"
     labels:

.github/workflows/mypy.yml

@@ -8,6 +8,7 @@ on:
   pull_request:
     paths:
       - "Tools/clinic/**"
+      - "Tools/cases_generator/**"
       - ".github/workflows/mypy.yml"
   workflow_dispatch:
 
@@ -25,15 +26,18 @@ concurrency:
 
 jobs:
   mypy:
-    name: Run mypy on Tools/clinic/
+    strategy:
+      matrix:
+        target: ["Tools/cases_generator", "Tools/clinic"]
+    name: Run mypy on ${{ matrix.target }}
     runs-on: ubuntu-latest
     timeout-minutes: 10
     steps:
       - uses: actions/checkout@v3
       - uses: actions/setup-python@v4
         with:
-          python-version: "3.x"
+          python-version: "3.11"
           cache: pip
-          cache-dependency-path: Tools/clinic/requirements-dev.txt
+          cache-dependency-path: Tools/requirements-dev.txt
-      - run: pip install -r Tools/clinic/requirements-dev.txt
+      - run: pip install -r Tools/requirements-dev.txt
-      - run: mypy --config-file Tools/clinic/mypy.ini
+      - run: mypy --config-file ${{ matrix.target }}/mypy.ini

Tools/cases_generator/flags.py

@@ -16,12 +16,11 @@ class InstructionFlags:
     HAS_FREE_FLAG: bool
     HAS_LOCAL_FLAG: bool
 
-    def __post_init__(self):
+    def __post_init__(self) -> None:
         self.bitmask = {name: (1 << i) for i, name in enumerate(self.names())}
 
     @staticmethod
-    def fromInstruction(instr: parsing.Node):
+    def fromInstruction(instr: parsing.Node) -> "InstructionFlags":
-
         has_free = (
             variable_used(instr, "PyCell_New")
             or variable_used(instr, "PyCell_GET")
@@ -41,7 +40,7 @@ class InstructionFlags:
         )
 
     @staticmethod
-    def newEmpty():
+    def newEmpty() -> "InstructionFlags":
         return InstructionFlags(False, False, False, False, False, False)
 
     def add(self, other: "InstructionFlags") -> None:
@@ -49,7 +48,7 @@ class InstructionFlags:
             if value:
                 setattr(self, name, value)
 
-    def names(self, value=None) -> list[str]:
+    def names(self, value: bool | None = None) -> list[str]:
         if value is None:
             return list(dataclasses.asdict(self).keys())
         return [n for n, v in dataclasses.asdict(self).items() if v == value]
@@ -62,7 +61,7 @@ class InstructionFlags:
         return flags
 
     @classmethod
-    def emit_macros(cls, out: Formatter):
+    def emit_macros(cls, out: Formatter) -> None:
         flags = cls.newEmpty()
         for name, value in flags.bitmask.items():
             out.emit(f"#define {name} ({value})")
@@ -90,9 +89,9 @@ def variable_used_unspecialized(node: parsing.Node, name: str) -> bool:
             text = "".join(token.text.split())
             # TODO: Handle nested #if
             if text == "#if":
-                if (
+                if i + 1 < len(node.tokens) and node.tokens[i + 1].text in (
-                    i + 1 < len(node.tokens)
+                    "ENABLE_SPECIALIZATION",
-                    and node.tokens[i + 1].text in ("ENABLE_SPECIALIZATION", "TIER_ONE")
+                    "TIER_ONE",
                 ):
                     skipping = True
             elif text in ("#else", "#endif"):

Tools/cases_generator/formatting.py

@@ -1,6 +1,7 @@
 import contextlib
 import re
 import typing
+from collections.abc import Iterator
 
 from parsing import StackEffect, Family
 
@@ -58,13 +59,13 @@ class Formatter:
             self.set_lineno(self.lineno + 1, self.filename)
 
     @contextlib.contextmanager
-    def indent(self):
+    def indent(self) -> Iterator[None]:
         self.prefix += "    "
         yield
         self.prefix = self.prefix[:-4]
 
     @contextlib.contextmanager
-    def block(self, head: str, tail: str = ""):
+    def block(self, head: str, tail: str = "") -> Iterator[None]:
         if head:
             self.emit(head + " {")
         else:
@@ -77,7 +78,7 @@ class Formatter:
         self,
         input_effects: list[StackEffect],
         output_effects: list[StackEffect],
-    ):
+    ) -> None:
         shrink, isym = list_effect_size(input_effects)
         grow, osym = list_effect_size(output_effects)
         diff = grow - shrink
@@ -90,7 +91,7 @@ class Formatter:
         if osym and osym != isym:
             self.emit(f"STACK_GROW({osym});")
 
-    def declare(self, dst: StackEffect, src: StackEffect | None):
+    def declare(self, dst: StackEffect, src: StackEffect | None) -> None:
         if dst.name == UNUSED or dst.cond == "0":
             return
         typ = f"{dst.type}" if dst.type else "PyObject *"
@@ -107,7 +108,7 @@ class Formatter:
         sepa = "" if typ.endswith("*") else " "
         self.emit(f"{typ}{sepa}{dst.name}{init};")
 
-    def assign(self, dst: StackEffect, src: StackEffect):
+    def assign(self, dst: StackEffect, src: StackEffect) -> None:
         if src.name == UNUSED or dst.name == UNUSED:
             return
         cast = self.cast(dst, src)

Tools/cases_generator/generate_cases.py

@@ -10,6 +10,7 @@ import os
 import posixpath
 import sys
 import typing
+from collections.abc import Iterator
 
 import stacking  # Early import to avoid circular import
 from analysis import Analyzer
@@ -23,7 +24,6 @@ from instructions import (
     MacroInstruction,
     MacroParts,
     PseudoInstruction,
-    StackEffect,
     OverriddenInstructionPlaceHolder,
     TIER_ONE,
     TIER_TWO,
@@ -80,12 +80,10 @@ SPECIALLY_HANDLED_ABSTRACT_INSTR = {
     "STORE_FAST",
     "STORE_FAST_MAYBE_NULL",
     "COPY",
-
     # Arithmetic
     "_BINARY_OP_MULTIPLY_INT",
     "_BINARY_OP_ADD_INT",
     "_BINARY_OP_SUBTRACT_INT",
-
 }
 
 arg_parser = argparse.ArgumentParser(
@@ -144,6 +142,7 @@ arg_parser.add_argument(
     default=DEFAULT_ABSTRACT_INTERPRETER_OUTPUT,
 )
 
+
 class Generator(Analyzer):
     def get_stack_effect_info(
         self, thing: parsing.InstDef | parsing.Macro | parsing.Pseudo
@@ -183,7 +182,8 @@ class Generator(Analyzer):
                     assert target_instr
                     target_popped = effect_str(target_instr.input_effects)
                     target_pushed = effect_str(target_instr.output_effects)
-                    if popped is None and pushed is None:
+                    if pushed is None:
+                        assert popped is None
                         popped, pushed = target_popped, target_pushed
                     else:
                         assert popped == target_popped
@@ -193,7 +193,7 @@ class Generator(Analyzer):
         return instr, popped, pushed
 
     @contextlib.contextmanager
-    def metadata_item(self, signature, open, close):
+    def metadata_item(self, signature: str, open: str, close: str) -> Iterator[None]:
         self.out.emit("")
         self.out.emit(f"extern {signature};")
         self.out.emit("#ifdef NEED_OPCODE_METADATA")
@@ -216,9 +216,10 @@ class Generator(Analyzer):
         def write_function(
             direction: str, data: list[tuple[AnyInstruction, str]]
         ) -> None:
-
             with self.metadata_item(
-                f"int _PyOpcode_num_{direction}(int opcode, int oparg, bool jump)", "", ""
+                f"int _PyOpcode_num_{direction}(int opcode, int oparg, bool jump)",
+                "",
+                "",
             ):
                 with self.out.block("switch(opcode)"):
                     for instr, effect in data:
@@ -243,23 +244,24 @@ class Generator(Analyzer):
         paths = f"\n{self.out.comment}   ".join(filenames)
         return f"{self.out.comment} from:\n{self.out.comment}   {paths}\n"
 
-    def write_provenance_header(self):
+    def write_provenance_header(self) -> None:
         self.out.write_raw(f"{self.out.comment} This file is generated by {THIS}\n")
         self.out.write_raw(self.from_source_files())
         self.out.write_raw(f"{self.out.comment} Do not edit!\n")
 
-    def assign_opcode_ids(self):
+    def assign_opcode_ids(self) -> None:
         """Assign IDs to opcodes"""
 
-        ops: list[(bool, str)] = []  # (has_arg, name) for each opcode
+        ops: list[tuple[bool, str]] = []  # (has_arg, name) for each opcode
         instrumented_ops: list[str] = []
 
         for instr in itertools.chain(
             [instr for instr in self.instrs.values() if instr.kind != "op"],
-            self.macro_instrs.values()):
+            self.macro_instrs.values(),
-
+        ):
+            assert isinstance(instr, (Instruction, MacroInstruction, PseudoInstruction))
             name = instr.name
-            if name.startswith('INSTRUMENTED_'):
+            if name.startswith("INSTRUMENTED_"):
                 instrumented_ops.append(name)
             else:
                 ops.append((instr.instr_flags.HAS_ARG_FLAG, name))
@@ -268,33 +270,32 @@ class Generator(Analyzer):
         # rather than bytecodes.c, so we need to add it explicitly
         # here (at least until we add something to bytecodes.c to
         # declare external instructions).
-        instrumented_ops.append('INSTRUMENTED_LINE')
+        instrumented_ops.append("INSTRUMENTED_LINE")
 
         # assert lists are unique
         assert len(set(ops)) == len(ops)
         assert len(set(instrumented_ops)) == len(instrumented_ops)
 
-        opname: list[str or None] = [None] * 512
+        opname: list[str | None] = [None] * 512
-        opmap: dict = {}
+        opmap: dict[str, int] = {}
-        markers: dict = {}
+        markers: dict[str, int] = {}
 
-        def map_op(op, name):
+        def map_op(op: int, name: str) -> None:
             assert op < len(opname)
             assert opname[op] is None
             assert name not in opmap
             opname[op] = name
             opmap[name] = op
 
-
         # 0 is reserved for cache entries. This helps debugging.
-        map_op(0, 'CACHE')
+        map_op(0, "CACHE")
 
         # 17 is reserved as it is the initial value for the specializing counter.
         # This helps catch cases where we attempt to execute a cache.
-        map_op(17, 'RESERVED')
+        map_op(17, "RESERVED")
 
         # 166 is RESUME - it is hard coded as such in Tools/build/deepfreeze.py
-        map_op(166, 'RESUME')
+        map_op(166, "RESUME")
 
         next_opcode = 1
 
@@ -306,13 +307,13 @@ class Generator(Analyzer):
             assert next_opcode < 255
             map_op(next_opcode, name)
 
-            if has_arg and 'HAVE_ARGUMENT' not in markers:
+            if has_arg and "HAVE_ARGUMENT" not in markers:
-                markers['HAVE_ARGUMENT'] = next_opcode
+                markers["HAVE_ARGUMENT"] = next_opcode
 
         # Instrumented opcodes are at the end of the valid range
         min_instrumented = 254 - (len(instrumented_ops) - 1)
         assert next_opcode <= min_instrumented
-        markers['MIN_INSTRUMENTED_OPCODE'] = min_instrumented
+        markers["MIN_INSTRUMENTED_OPCODE"] = min_instrumented
         for i, op in enumerate(instrumented_ops):
             map_op(min_instrumented + i, op)
 
@@ -320,11 +321,13 @@ class Generator(Analyzer):
         for i, op in enumerate(sorted(self.pseudos)):
             map_op(256 + i, op)
 
-        assert 255 not in opmap  # 255 is reserved
+        assert 255 not in opmap.values()  # 255 is reserved
         self.opmap = opmap
         self.markers = markers
 
-    def write_opcode_ids(self, opcode_ids_h_filename, opcode_targets_filename):
+    def write_opcode_ids(
+        self, opcode_ids_h_filename: str, opcode_targets_filename: str
+    ) -> None:
         """Write header file that defined the opcode IDs"""
 
         with open(opcode_ids_h_filename, "w") as f:
@@ -337,15 +340,15 @@ class Generator(Analyzer):
             self.out.emit("#ifndef Py_OPCODE_IDS_H")
             self.out.emit("#define Py_OPCODE_IDS_H")
             self.out.emit("#ifdef __cplusplus")
-            self.out.emit("extern \"C\" {")
+            self.out.emit('extern "C" {')
             self.out.emit("#endif")
             self.out.emit("")
             self.out.emit("/* Instruction opcodes for compiled code */")
 
-            def define(name, opcode):
+            def define(name: str, opcode: int) -> None:
                 self.out.emit(f"#define {name:<38} {opcode:>3}")
 
-            all_pairs = []
+            all_pairs: list[tuple[int, int, str]] = []
             # the second item in the tuple sorts the markers before the ops
             all_pairs.extend((i, 1, name) for (name, i) in self.markers.items())
             all_pairs.extend((i, 2, name) for (name, i) in self.opmap.items())
@@ -370,7 +373,6 @@ class Generator(Analyzer):
                         targets[op] = f"TARGET_{name}"
                 f.write(",\n".join([f"    &&{s}" for s in targets]))
 
-
     def write_metadata(self, metadata_filename: str, pymetadata_filename: str) -> None:
         """Write instruction metadata to output file."""
 
@@ -469,7 +471,7 @@ class Generator(Analyzer):
                 "const struct opcode_metadata "
                 "_PyOpcode_opcode_metadata[OPCODE_METADATA_SIZE]",
                 "=",
-                ";"
+                ";",
             ):
                 # Write metadata for each instruction
                 for thing in self.everything:
@@ -482,7 +484,9 @@ class Generator(Analyzer):
                         case parsing.Macro():
                             self.write_metadata_for_macro(self.macro_instrs[thing.name])
                         case parsing.Pseudo():
-                            self.write_metadata_for_pseudo(self.pseudo_instrs[thing.name])
+                            self.write_metadata_for_pseudo(
+                                self.pseudo_instrs[thing.name]
+                            )
                         case _:
                             typing.assert_never(thing)
 
@@ -490,7 +494,7 @@ class Generator(Analyzer):
                 "const struct opcode_macro_expansion "
                 "_PyOpcode_macro_expansion[OPCODE_MACRO_EXPANSION_SIZE]",
                 "=",
-                ";"
+                ";",
             ):
                 # Write macro expansion for each non-pseudo instruction
                 for thing in self.everything:
@@ -529,7 +533,9 @@ class Generator(Analyzer):
                 self.write_uop_items(lambda name, counter: f'[{name}] = "{name}",')
 
             with self.metadata_item(
-                f"const char *const _PyOpcode_OpName[{1 + max(self.opmap.values())}]", "=", ";"
+                f"const char *const _PyOpcode_OpName[{1 + max(self.opmap.values())}]",
+                "=",
+                ";",
             ):
                 for name in self.opmap:
                     self.out.emit(f'[{name}] = "{name}",')
@@ -542,11 +548,9 @@ class Generator(Analyzer):
                 for m in family.members:
                     deoptcodes[m] = name
             # special case:
-            deoptcodes['BINARY_OP_INPLACE_ADD_UNICODE'] = 'BINARY_OP'
+            deoptcodes["BINARY_OP_INPLACE_ADD_UNICODE"] = "BINARY_OP"
 
-            with self.metadata_item(
+            with self.metadata_item(f"const uint8_t _PyOpcode_Deopt[256]", "=", ";"):
-                f"const uint8_t _PyOpcode_Deopt[256]", "=", ";"
-            ):
                 for opt, deopt in sorted(deoptcodes.items()):
                     self.out.emit(f"[{opt}] = {deopt},")
 
@@ -604,10 +608,9 @@ class Generator(Analyzer):
                     if name not in specialized_ops:
                         self.out.emit(f"'{name}': {op},")
 
-            for name in ['MIN_INSTRUMENTED_OPCODE', 'HAVE_ARGUMENT']:
+            for name in ["MIN_INSTRUMENTED_OPCODE", "HAVE_ARGUMENT"]:
                 self.out.emit(f"{name} = {self.markers[name]}")
 
-
     def write_pseudo_instrs(self) -> None:
         """Write the IS_PSEUDO_INSTR macro"""
         self.out.emit("\n\n#define IS_PSEUDO_INSTR(OP)  ( \\")
@@ -834,7 +837,10 @@ class Generator(Analyzer):
                         pass
                     case parsing.InstDef():
                         instr = AbstractInstruction(self.instrs[thing.name].inst)
-                        if instr.is_viable_uop() and instr.name not in SPECIALLY_HANDLED_ABSTRACT_INSTR:
+                        if (
+                            instr.is_viable_uop()
+                            and instr.name not in SPECIALLY_HANDLED_ABSTRACT_INSTR
+                        ):
                             self.out.emit("")
                             with self.out.block(f"case {thing.name}:"):
                                 instr.write(self.out, tier=TIER_TWO)
@@ -878,7 +884,7 @@ class Generator(Analyzer):
                 self.out.emit(f"DISPATCH();")
 
 
-def main():
+def main() -> None:
     """Parse command line, parse input, analyze, write output."""
     args = arg_parser.parse_args()  # Prints message and sys.exit(2) on error
     if len(args.input) == 0:
@@ -899,8 +905,9 @@ def main():
     a.write_opcode_ids(args.opcode_ids_h, args.opcode_targets_h)
     a.write_metadata(args.metadata, args.pymetadata)
     a.write_executor_instructions(args.executor_cases, args.emit_line_directives)
-    a.write_abstract_interpreter_instructions(args.abstract_interpreter_cases,
+    a.write_abstract_interpreter_instructions(
-                                              args.emit_line_directives)
+        args.abstract_interpreter_cases, args.emit_line_directives
+    )
 
 
 if __name__ == "__main__":

Tools/cases_generator/lexer.py

@@ -4,132 +4,221 @@
 
 import re
 from dataclasses import dataclass
+from collections.abc import Iterator
 
-def choice(*opts):
+
+def choice(*opts: str) -> str:
     return "|".join("(%s)" % opt for opt in opts)
 
+
 # Regexes
 
 # Longer operators must go before shorter ones.
 
-PLUSPLUS = r'\+\+'
+PLUSPLUS = r"\+\+"
-MINUSMINUS = r'--'
+MINUSMINUS = r"--"
 
 # ->
-ARROW = r'->'
+ARROW = r"->"
-ELLIPSIS = r'\.\.\.'
+ELLIPSIS = r"\.\.\."
 
 # Assignment operators
-TIMESEQUAL = r'\*='
+TIMESEQUAL = r"\*="
-DIVEQUAL = r'/='
+DIVEQUAL = r"/="
-MODEQUAL = r'%='
+MODEQUAL = r"%="
-PLUSEQUAL = r'\+='
+PLUSEQUAL = r"\+="
-MINUSEQUAL = r'-='
+MINUSEQUAL = r"-="
-LSHIFTEQUAL = r'<<='
+LSHIFTEQUAL = r"<<="
-RSHIFTEQUAL = r'>>='
+RSHIFTEQUAL = r">>="
-ANDEQUAL = r'&='
+ANDEQUAL = r"&="
-OREQUAL = r'\|='
+OREQUAL = r"\|="
-XOREQUAL = r'\^='
+XOREQUAL = r"\^="
 
 # Operators
-PLUS = r'\+'
+PLUS = r"\+"
-MINUS = r'-'
+MINUS = r"-"
-TIMES = r'\*'
+TIMES = r"\*"
-DIVIDE = r'/'
+DIVIDE = r"/"
-MOD = r'%'
+MOD = r"%"
-NOT = r'~'
+NOT = r"~"
-XOR = r'\^'
+XOR = r"\^"
-LOR = r'\|\|'
+LOR = r"\|\|"
-LAND = r'&&'
+LAND = r"&&"
-LSHIFT = r'<<'
+LSHIFT = r"<<"
-RSHIFT = r'>>'
+RSHIFT = r">>"
-LE = r'<='
+LE = r"<="
-GE = r'>='
+GE = r">="
-EQ = r'=='
+EQ = r"=="
-NE = r'!='
+NE = r"!="
-LT = r'<'
+LT = r"<"
-GT = r'>'
+GT = r">"
-LNOT = r'!'
+LNOT = r"!"
-OR = r'\|'
+OR = r"\|"
-AND = r'&'
+AND = r"&"
-EQUALS = r'='
+EQUALS = r"="
 
 # ?
-CONDOP = r'\?'
+CONDOP = r"\?"
 
 # Delimiters
-LPAREN = r'\('
+LPAREN = r"\("
-RPAREN = r'\)'
+RPAREN = r"\)"
-LBRACKET = r'\['
+LBRACKET = r"\["
-RBRACKET = r'\]'
+RBRACKET = r"\]"
-LBRACE = r'\{'
+LBRACE = r"\{"
-RBRACE = r'\}'
+RBRACE = r"\}"
-COMMA = r','
+COMMA = r","
-PERIOD = r'\.'
+PERIOD = r"\."
-SEMI = r';'
+SEMI = r";"
-COLON = r':'
+COLON = r":"
-BACKSLASH = r'\\'
+BACKSLASH = r"\\"
 
 operators = {op: pattern for op, pattern in globals().items() if op == op.upper()}
 for op in operators:
     globals()[op] = op
-opmap = { pattern.replace("\\", "") or '\\' : op for op, pattern in operators.items() }
+opmap = {pattern.replace("\\", "") or "\\": op for op, pattern in operators.items()}
 
 # Macros
-macro = r'# *(ifdef|ifndef|undef|define|error|endif|if|else|include|#)'
+macro = r"# *(ifdef|ifndef|undef|define|error|endif|if|else|include|#)"
-MACRO = 'MACRO'
+MACRO = "MACRO"
 
-id_re = r'[a-zA-Z_][0-9a-zA-Z_]*'
+id_re = r"[a-zA-Z_][0-9a-zA-Z_]*"
-IDENTIFIER = 'IDENTIFIER'
+IDENTIFIER = "IDENTIFIER"
 
-suffix = r'([uU]?[lL]?[lL]?)'
+suffix = r"([uU]?[lL]?[lL]?)"
-octal = r'0[0-7]+' + suffix
+octal = r"0[0-7]+" + suffix
-hex = r'0[xX][0-9a-fA-F]+'
+hex = r"0[xX][0-9a-fA-F]+"
-decimal_digits = r'(0|[1-9][0-9]*)'
+decimal_digits = r"(0|[1-9][0-9]*)"
 decimal = decimal_digits + suffix
 
 
 exponent = r"""([eE][-+]?[0-9]+)"""
 fraction = r"""([0-9]*\.[0-9]+)|([0-9]+\.)"""
-float = '(((('+fraction+')'+exponent+'?)|([0-9]+'+exponent+'))[FfLl]?)'
+float = "((((" + fraction + ")" + exponent + "?)|([0-9]+" + exponent + "))[FfLl]?)"
 
 number_re = choice(octal, hex, float, decimal)
-NUMBER = 'NUMBER'
+NUMBER = "NUMBER"
 
 simple_escape = r"""([a-zA-Z._~!=&\^\-\\?'"])"""
 decimal_escape = r"""(\d+)"""
 hex_escape = r"""(x[0-9a-fA-F]+)"""
-escape_sequence = r"""(\\("""+simple_escape+'|'+decimal_escape+'|'+hex_escape+'))'
+escape_sequence = (
-string_char = r"""([^"\\\n]|"""+escape_sequence+')'
+    r"""(\\(""" + simple_escape + "|" + decimal_escape + "|" + hex_escape + "))"
+)
+string_char = r"""([^"\\\n]|""" + escape_sequence + ")"
 str_re = '"' + string_char + '*"'
-STRING = 'STRING'
+STRING = "STRING"
-char = r'\'.\''  # TODO: escape sequence
+char = r"\'.\'"  # TODO: escape sequence
-CHARACTER = 'CHARACTER'
+CHARACTER = "CHARACTER"
 
-comment_re = r'//.*|/\*([^*]|\*[^/])*\*/'
+comment_re = r"//.*|/\*([^*]|\*[^/])*\*/"
-COMMENT = 'COMMENT'
+COMMENT = "COMMENT"
 
 newline = r"\n"
-invalid = r"\S"  # A single non-space character that's not caught by any of the other patterns
+invalid = (
-matcher = re.compile(choice(id_re, number_re, str_re, char, newline, macro, comment_re, *operators.values(), invalid))
+    r"\S"  # A single non-space character that's not caught by any of the other patterns
-letter = re.compile(r'[a-zA-Z_]')
-
-kwds = (
-    'AUTO', 'BREAK', 'CASE', 'CHAR', 'CONST',
-    'CONTINUE', 'DEFAULT', 'DO', 'DOUBLE', 'ELSE', 'ENUM', 'EXTERN',
-    'FLOAT', 'FOR', 'GOTO', 'IF', 'INLINE', 'INT', 'LONG', 'OVERRIDE',
-    'REGISTER', 'OFFSETOF',
-    'RESTRICT', 'RETURN', 'SHORT', 'SIGNED', 'SIZEOF', 'STATIC', 'STRUCT',
-    'SWITCH', 'TYPEDEF', 'UNION', 'UNSIGNED', 'VOID',
-    'VOLATILE', 'WHILE'
 )
-for name in kwds:
+matcher = re.compile(
-    globals()[name] = name
+    choice(
+        id_re,
+        number_re,
+        str_re,
+        char,
+        newline,
+        macro,
+        comment_re,
+        *operators.values(),
+        invalid,
+    )
+)
+letter = re.compile(r"[a-zA-Z_]")
+
+
+kwds = []
+AUTO = "AUTO"
+kwds.append(AUTO)
+BREAK = "BREAK"
+kwds.append(BREAK)
+CASE = "CASE"
+kwds.append(CASE)
+CHAR = "CHAR"
+kwds.append(CHAR)
+CONST = "CONST"
+kwds.append(CONST)
+CONTINUE = "CONTINUE"
+kwds.append(CONTINUE)
+DEFAULT = "DEFAULT"
+kwds.append(DEFAULT)
+DO = "DO"
+kwds.append(DO)
+DOUBLE = "DOUBLE"
+kwds.append(DOUBLE)
+ELSE = "ELSE"
+kwds.append(ELSE)
+ENUM = "ENUM"
+kwds.append(ENUM)
+EXTERN = "EXTERN"
+kwds.append(EXTERN)
+FLOAT = "FLOAT"
+kwds.append(FLOAT)
+FOR = "FOR"
+kwds.append(FOR)
+GOTO = "GOTO"
+kwds.append(GOTO)
+IF = "IF"
+kwds.append(IF)
+INLINE = "INLINE"
+kwds.append(INLINE)
+INT = "INT"
+kwds.append(INT)
+LONG = "LONG"
+kwds.append(LONG)
+OVERRIDE = "OVERRIDE"
+kwds.append(OVERRIDE)
+REGISTER = "REGISTER"
+kwds.append(REGISTER)
+OFFSETOF = "OFFSETOF"
+kwds.append(OFFSETOF)
+RESTRICT = "RESTRICT"
+kwds.append(RESTRICT)
+RETURN = "RETURN"
+kwds.append(RETURN)
+SHORT = "SHORT"
+kwds.append(SHORT)
+SIGNED = "SIGNED"
+kwds.append(SIGNED)
+SIZEOF = "SIZEOF"
+kwds.append(SIZEOF)
+STATIC = "STATIC"
+kwds.append(STATIC)
+STRUCT = "STRUCT"
+kwds.append(STRUCT)
+SWITCH = "SWITCH"
+kwds.append(SWITCH)
+TYPEDEF = "TYPEDEF"
+kwds.append(TYPEDEF)
+UNION = "UNION"
+kwds.append(UNION)
+UNSIGNED = "UNSIGNED"
+kwds.append(UNSIGNED)
+VOID = "VOID"
+kwds.append(VOID)
+VOLATILE = "VOLATILE"
+kwds.append(VOLATILE)
+WHILE = "WHILE"
+kwds.append(WHILE)
 keywords = {name.lower(): name for name in kwds}
 
+__all__ = []
+__all__.extend(kwds)
+
 
 def make_syntax_error(
-    message: str, filename: str, line: int, column: int, line_text: str,
+    message: str,
+    filename: str | None,
+    line: int,
+    column: int,
+    line_text: str,
 ) -> SyntaxError:
     return SyntaxError(message, (filename, line, column, line_text))
 
@@ -142,30 +231,30 @@ class Token:
     end: tuple[int, int]
 
     @property
-    def line(self):
+    def line(self) -> int:
         return self.begin[0]
 
     @property
-    def column(self):
+    def column(self) -> int:
         return self.begin[1]
 
     @property
-    def end_line(self):
+    def end_line(self) -> int:
         return self.end[0]
 
     @property
-    def end_column(self):
+    def end_column(self) -> int:
         return self.end[1]
 
     @property
-    def width(self):
+    def width(self) -> int:
         return self.end[1] - self.begin[1]
 
-    def replaceText(self, txt):
+    def replaceText(self, txt: str) -> "Token":
         assert isinstance(txt, str)
         return Token(self.kind, txt, self.begin, self.end)
 
-    def __repr__(self):
+    def __repr__(self) -> str:
         b0, b1 = self.begin
         e0, e1 = self.end
         if b0 == e0:
@@ -174,7 +263,7 @@ class Token:
             return f"{self.kind}({self.text!r}, {b0}:{b1}, {e0}:{e1})"
 
 
-def tokenize(src, line=1, filename=None):
+def tokenize(src: str, line: int = 1, filename: str | None = None) -> Iterator[Token]:
     linestart = -1
     for m in matcher.finditer(src):
         start, end = m.span()
@@ -183,37 +272,42 @@ def tokenize(src, line=1, filename=None):
             kind = keywords[text]
         elif letter.match(text):
             kind = IDENTIFIER
-        elif text == '...':
+        elif text == "...":
             kind = ELLIPSIS
-        elif text == '.':
+        elif text == ".":
             kind = PERIOD
-        elif text[0] in '0123456789.':
+        elif text[0] in "0123456789.":
             kind = NUMBER
         elif text[0] == '"':
             kind = STRING
         elif text in opmap:
             kind = opmap[text]
-        elif text == '\n':
+        elif text == "\n":
             linestart = start
             line += 1
-            kind = '\n'
+            kind = "\n"
         elif text[0] == "'":
             kind = CHARACTER
-        elif text[0] == '#':
+        elif text[0] == "#":
             kind = MACRO
-        elif text[0] == '/' and text[1] in '/*':
+        elif text[0] == "/" and text[1] in "/*":
             kind = COMMENT
         else:
             lineend = src.find("\n", start)
             if lineend == -1:
                 lineend = len(src)
-            raise make_syntax_error(f"Bad token: {text}",
+            raise make_syntax_error(
-                filename, line, start-linestart+1, src[linestart:lineend])
+                f"Bad token: {text}",
+                filename,
+                line,
+                start - linestart + 1,
+                src[linestart:lineend],
+            )
         if kind == COMMENT:
             begin = line, start - linestart
-            newlines = text.count('\n')
+            newlines = text.count("\n")
             if newlines:
-                linestart = start + text.rfind('\n')
+                linestart = start + text.rfind("\n")
                 line += newlines
         else:
             begin = line, start - linestart
@@ -221,10 +315,6 @@ def tokenize(src, line=1, filename=None):
             yield Token(kind, text, begin, (line, start - linestart + len(text)))
 
 
-__all__ = []
-__all__.extend([kind for kind in globals() if kind.upper() == kind])
-
-
 def to_text(tkns: list[Token], dedent: int = 0) -> str:
     res: list[str] = []
     line, col = -1, 1 + dedent
@@ -235,21 +325,22 @@ def to_text(tkns: list[Token], dedent: int = 0) -> str:
         # assert(l >= line), (line, txt, start, end)
         while l > line:
             line += 1
-            res.append('\n')
+            res.append("\n")
             col = 1 + dedent
-        res.append(' '*(c-col))
+        res.append(" " * (c - col))
         text = tkn.text
-        if dedent != 0 and tkn.kind == 'COMMENT' and '\n' in text:
+        if dedent != 0 and tkn.kind == "COMMENT" and "\n" in text:
             if dedent < 0:
-                text = text.replace('\n', '\n' + ' '*-dedent)
+                text = text.replace("\n", "\n" + " " * -dedent)
             # TODO: dedent > 0
         res.append(text)
         line, col = tkn.end
-    return ''.join(res)
+    return "".join(res)
 
 
 if __name__ == "__main__":
     import sys
+
     filename = sys.argv[1]
     if filename == "-c":
         src = sys.argv[2]

Tools/cases_generator/mypy.ini

@@ -0,0 +1,14 @@
+[mypy]
+files = Tools/cases_generator/
+pretty = True
+
+python_version = 3.10
+
+# Be strict:
+strict = True
+strict_concatenate = True
+enable_error_code = ignore-without-code,redundant-expr,truthy-bool
+
+# Don't enable this one yet;
+# it has a lot of false positives on `cases_generator`
+warn_unreachable = False

Tools/cases_generator/parsing.py

@@ -32,7 +32,7 @@ class Context(NamedTuple):
     end: int
     owner: PLexer
 
-    def __repr__(self):
+    def __repr__(self) -> str:
         return f"<{self.owner.filename}: {self.begin}-{self.end}>"
 
 
@@ -75,7 +75,7 @@ class StackEffect(Node):
     size: str = ""  # Optional `[size]`
     # Note: size cannot be combined with type or cond
 
-    def __repr__(self):
+    def __repr__(self) -> str:
         items = [self.name, self.type, self.cond, self.size]
         while items and items[-1] == "":
             del items[-1]

Tools/cases_generator/stacking.py

@@ -413,22 +413,22 @@ def write_components(
     return next_instr_is_set
 
 
-def write_single_instr_for_abstract_interp(
+def write_single_instr_for_abstract_interp(instr: Instruction, out: Formatter) -> None:
-    instr: Instruction, out: Formatter
-):
     try:
         _write_components_for_abstract_interp(
             [Component(instr, instr.active_caches)],
             out,
         )
     except AssertionError as err:
-        raise AssertionError(f"Error writing abstract instruction {instr.name}") from err
+        raise AssertionError(
+            f"Error writing abstract instruction {instr.name}"
+        ) from err
 
 
 def _write_components_for_abstract_interp(
     parts: list[Component],
     out: Formatter,
-):
+) -> None:
     managers = get_managers(parts)
     for mgr in managers:
         if mgr is managers[-1]:
@@ -438,5 +438,7 @@ def _write_components_for_abstract_interp(
         # NULL out the output stack effects
         for poke in mgr.pokes:
             if not poke.effect.size and poke.effect.name not in mgr.instr.unmoved_names:
-                out.emit(f"PARTITIONNODE_OVERWRITE((_Py_PARTITIONNODE_t *)"
+                out.emit(
-                         f"PARTITIONNODE_NULLROOT, PEEK(-({poke.offset.as_index()})), true);")
+                    f"PARTITIONNODE_OVERWRITE((_Py_PARTITIONNODE_t *)"
+                    f"PARTITIONNODE_NULLROOT, PEEK(-({poke.offset.as_index()})), true);"
+                )

Tools/clinic/requirements-dev.txt

@@ -1,2 +0,0 @@
-# Requirements file for external linters and checks we run on Tools/clinic/ in CI
-mypy==1.4.1

Tools/requirements-dev.txt

@@ -0,0 +1,3 @@
+# Requirements file for external linters and checks we run on
+# Tools/clinic and Tools/cases_generator/ in CI
+mypy==1.5.1