bpo-45850: Implement deep-freeze on Windows (#29648)

Implement changes to build with deep-frozen modules on Windows.
Note that we now require Python 3.10 as the "bootstrap" or "host" Python.
This causes a modest startup speed (around 7%) on Windows.

1 files changed, 46 insertions, 12 deletions

diff --git a/Tools/scripts/deepfreeze.py b/Tools/scripts/deepfreeze.py
index 074127f9492..b6d52b74549 100644
--- a/Tools/scripts/deepfreeze.py
+++ b/Tools/scripts/deepfreeze.py
@@ -1,13 +1,16 @@
 import argparse
+import ast
 import builtins
 import collections
 import contextlib
 import os
-import sys
+import re
 import time
 import types
 import typing
 
+import umarshal
+
 verbose = False
 
 
@@ -55,7 +58,8 @@ def get_localsplus_counts(code: types.CodeType,
             nplaincellvars += 1
         elif kind & CO_FAST_FREE:
             nfreevars += 1
-    assert nlocals == len(code.co_varnames) == code.co_nlocals
+    assert nlocals == len(code.co_varnames) == code.co_nlocals, \
+        (nlocals, len(code.co_varnames), code.co_nlocals)
     assert ncellvars == len(code.co_cellvars)
     assert nfreevars == len(code.co_freevars)
     assert len(names) == nlocals + nplaincellvars + nfreevars
@@ -274,14 +278,7 @@ class Printer:
                             self.write(item + ",")
         return f"& {name}._object.ob_base.ob_base"
 
-    def generate_int(self, name: str, i: int) -> str:
-        maxint = sys.maxsize
-        if maxint == 2**31 - 1:
-            digit = 2**15
-        elif maxint == 2**63 - 1:
-            digit = 2**30
-        else:
-            assert False, f"What int size is this system?!? {maxint=}"
+    def _generate_int_for_bits(self, name: str, i: int, digit: int) -> None:
         sign = -1 if i < 0 else 0 if i == 0 else +1
         i = abs(i)
         digits: list[int] = []
@@ -298,6 +295,20 @@ class Printer:
             if digits:
                 ds = ", ".join(map(str, digits))
                 self.write(f".ob_digit = {{ {ds} }},")
+
+    def generate_int(self, name: str, i: int) -> str:
+        if abs(i) < 2**15:
+            self._generate_int_for_bits(name, i, 2**15)
+        else:
+            connective = "if"
+            for bits_in_digit in 15, 30:
+                self.write(f"#{connective} PYLONG_BITS_IN_DIGIT == {bits_in_digit}")
+                self._generate_int_for_bits(name, i, 2**bits_in_digit)
+                connective = "elif"
+            self.write("#else")
+            self.write('#error "PYLONG_BITS_IN_DIGIT should be 15 or 30"')
+            self.write("#endif")
+            # If neither clause applies, it won't compile
         return f"& {name}.ob_base.ob_base"
 
     def generate_float(self, name: str, x: float) -> str:
@@ -326,7 +337,7 @@ class Printer:
             return self.cache[key]
         self.misses += 1
         match obj:
-            case types.CodeType() as code:
+            case types.CodeType() | umarshal.Code() as code:
                 val = self.generate_code(name, code)
             case tuple(t):
                 val = self.generate_tuple(name, t)
@@ -367,8 +378,31 @@ _Py_get_%%NAME%%_toplevel(void)
 }
 """
 
+FROZEN_COMMENT = "/* Auto-generated by Programs/_freeze_module.c */"
+
+FROZEN_DATA_LINE = r"\s*(\d+,\s*)+\s*"
+
+
+def is_frozen_header(source: str) -> bool:
+    return source.startswith(FROZEN_COMMENT)
+
+
+def decode_frozen_data(source: str) -> types.CodeType:
+    lines = source.splitlines()
+    while lines and re.match(FROZEN_DATA_LINE, lines[0]) is None:
+        del lines[0]
+    while lines and re.match(FROZEN_DATA_LINE, lines[-1]) is None:
+        del lines[-1]
+    values: tuple[int, ...] = ast.literal_eval("".join(lines))
+    data = bytes(values)
+    return umarshal.loads(data)
+
+
 def generate(source: str, filename: str, modname: str, file: typing.TextIO) -> None:
-    code = compile(source, filename, "exec")
+    if is_frozen_header(source):
+        code = decode_frozen_data(source)
+    else:
+        code = compile(source, filename, "exec")
     printer = Printer(file)
     printer.generate("toplevel", code)
     printer.write("")