cpython/Tools/build/generate_global_objects.py
Petr Viktorin 6f1d448bc1
gh-113993: Allow interned strings to be mortal, and fix related issues (GH-120520)
* Add an InternalDocs file describing how interning should work and how to use it.

* Add internal functions to *explicitly* request what kind of interning is done:
  - `_PyUnicode_InternMortal`
  - `_PyUnicode_InternImmortal`
  - `_PyUnicode_InternStatic`

* Switch uses of `PyUnicode_InternInPlace` to those.

* Disallow using `_Py_SetImmortal` on strings directly.
  You should use `_PyUnicode_InternImmortal` instead:
  - Strings should be interned before immortalization, otherwise you're possibly
    interning a immortalizing copy.
  - `_Py_SetImmortal` doesn't handle the `SSTATE_INTERNED_MORTAL` to
    `SSTATE_INTERNED_IMMORTAL` update, and those flags can't be changed in
    backports, as they are now part of public API and version-specific ABI.

* Add private `_only_immortal` argument for `sys.getunicodeinternedsize`, used in refleak test machinery.

* Make sure the statically allocated string singletons are unique. This means these sets are now disjoint:
  - `_Py_ID`
  - `_Py_STR` (including the empty string)
  - one-character latin-1 singletons

  Now, when you intern a singleton, that exact singleton will be interned.

* Add a `_Py_LATIN1_CHR` macro, use it instead of `_Py_ID`/`_Py_STR` for one-character latin-1 singletons everywhere (including Clinic).

* Intern `_Py_STR` singletons at startup.

* For free-threaded builds, intern `_Py_LATIN1_CHR` singletons at startup.

* Beef up the tests. Cover internal details (marked with `@cpython_only`).

* Add lots of assertions

Co-Authored-By: Eric Snow <ericsnowcurrently@gmail.com>
2024-06-21 17:19:31 +02:00

464 lines
15 KiB
Python

import contextlib
import io
import os.path
import re
SCRIPT_NAME = 'Tools/build/generate_global_objects.py'
__file__ = os.path.abspath(__file__)
ROOT = os.path.dirname(os.path.dirname(os.path.dirname(__file__)))
INTERNAL = os.path.join(ROOT, 'Include', 'internal')
IGNORED = {
'ACTION', # Python/_warnings.c
'ATTR', # Python/_warnings.c and Objects/funcobject.c
'DUNDER', # Objects/typeobject.c
'RDUNDER', # Objects/typeobject.c
'SPECIAL', # Objects/weakrefobject.c
'NAME', # Objects/typeobject.c
}
IDENTIFIERS = [
# from ADD() Python/_warnings.c
'default',
'ignore',
# from GET_WARNINGS_ATTR() in Python/_warnings.c
'WarningMessage',
'_showwarnmsg',
'_warn_unawaited_coroutine',
'defaultaction',
'filters',
'onceregistry',
# from WRAP_METHOD() in Objects/weakrefobject.c
'__bytes__',
'__reversed__',
# from COPY_ATTR() in Objects/funcobject.c
'__module__',
'__name__',
'__qualname__',
'__doc__',
'__annotations__',
# from SLOT* in Objects/typeobject.c
'__abs__',
'__add__',
'__aiter__',
'__and__',
'__anext__',
'__await__',
'__bool__',
'__call__',
'__contains__',
'__del__',
'__delattr__',
'__delete__',
'__delitem__',
'__eq__',
'__float__',
'__floordiv__',
'__ge__',
'__get__',
'__getattr__',
'__getattribute__',
'__getitem__',
'__gt__',
'__hash__',
'__iadd__',
'__iand__',
'__ifloordiv__',
'__ilshift__',
'__imatmul__',
'__imod__',
'__imul__',
'__index__',
'__init__',
'__int__',
'__invert__',
'__ior__',
'__ipow__',
'__irshift__',
'__isub__',
'__iter__',
'__itruediv__',
'__ixor__',
'__le__',
'__len__',
'__lshift__',
'__lt__',
'__matmul__',
'__mod__',
'__mul__',
'__ne__',
'__neg__',
'__new__',
'__next__',
'__or__',
'__pos__',
'__pow__',
'__radd__',
'__rand__',
'__repr__',
'__rfloordiv__',
'__rlshift__',
'__rmatmul__',
'__rmod__',
'__rmul__',
'__ror__',
'__rpow__',
'__rrshift__',
'__rshift__',
'__rsub__',
'__rtruediv__',
'__rxor__',
'__set__',
'__setattr__',
'__setitem__',
'__str__',
'__sub__',
'__truediv__',
'__xor__',
'__divmod__',
'__rdivmod__',
'__buffer__',
'__release_buffer__',
#Workarounds for GH-108918
'alias',
'args',
'exc_type',
'exc_value',
'self',
'traceback',
]
NON_GENERATED_IMMORTAL_OBJECTS = [
# The generated ones come from generate_runtime_init().
'(PyObject *)&_Py_SINGLETON(bytes_empty)',
'(PyObject *)&_Py_SINGLETON(tuple_empty)',
'(PyObject *)&_Py_SINGLETON(hamt_bitmap_node_empty)',
'(PyObject *)&_Py_INTERP_SINGLETON(interp, hamt_empty)',
'(PyObject *)&_Py_SINGLETON(context_token_missing)',
]
#######################################
# helpers
def iter_files():
for name in ('Modules', 'Objects', 'Parser', 'PC', 'Programs', 'Python'):
root = os.path.join(ROOT, name)
for dirname, _, files in os.walk(root):
for name in files:
if not name.endswith(('.c', '.h')):
continue
yield os.path.join(dirname, name)
def iter_global_strings():
id_regex = re.compile(r'\b_Py_ID\((\w+)\)')
str_regex = re.compile(r'\b_Py_DECLARE_STR\((\w+), "(.*?)"\)')
for filename in iter_files():
try:
infile = open(filename, encoding='utf-8')
except FileNotFoundError:
# The file must have been a temporary file.
continue
with infile:
for lno, line in enumerate(infile, 1):
for m in id_regex.finditer(line):
identifier, = m.groups()
yield identifier, None, filename, lno, line
for m in str_regex.finditer(line):
varname, string = m.groups()
yield varname, string, filename, lno, line
def iter_to_marker(lines, marker):
for line in lines:
if line.rstrip() == marker:
break
yield line
class Printer:
def __init__(self, file):
self.level = 0
self.file = file
self.continuation = [False]
@contextlib.contextmanager
def indent(self):
save_level = self.level
try:
self.level += 1
yield
finally:
self.level = save_level
def write(self, arg):
eol = '\n'
if self.continuation[-1]:
eol = f' \\{eol}' if arg else f'\\{eol}'
self.file.writelines((" "*self.level, arg, eol))
@contextlib.contextmanager
def block(self, prefix, suffix="", *, continuation=None):
if continuation is None:
continuation = self.continuation[-1]
self.continuation.append(continuation)
self.write(prefix + " {")
with self.indent():
yield
self.continuation.pop()
self.write("}" + suffix)
@contextlib.contextmanager
def open_for_changes(filename, orig):
"""Like open() but only write to the file if it changed."""
outfile = io.StringIO()
yield outfile
text = outfile.getvalue()
if text != orig:
with open(filename, 'w', encoding='utf-8') as outfile:
outfile.write(text)
else:
print(f'# not changed: {filename}')
#######################################
# the global objects
START = f'/* The following is auto-generated by {SCRIPT_NAME}. */'
END = '/* End auto-generated code */'
def generate_global_strings(identifiers, strings):
filename = os.path.join(INTERNAL, 'pycore_global_strings.h')
# Read the non-generated part of the file.
with open(filename) as infile:
orig = infile.read()
lines = iter(orig.rstrip().splitlines())
before = '\n'.join(iter_to_marker(lines, START))
for _ in iter_to_marker(lines, END):
pass
after = '\n'.join(lines)
# Generate the file.
with open_for_changes(filename, orig) as outfile:
printer = Printer(outfile)
printer.write(before)
printer.write(START)
with printer.block('struct _Py_global_strings', ';'):
with printer.block('struct', ' literals;'):
for literal, name in sorted(strings.items(), key=lambda x: x[1]):
printer.write(f'STRUCT_FOR_STR({name}, "{literal}")')
outfile.write('\n')
with printer.block('struct', ' identifiers;'):
for name in sorted(identifiers):
assert name.isidentifier(), name
printer.write(f'STRUCT_FOR_ID({name})')
with printer.block('struct', ' ascii[128];'):
printer.write("PyASCIIObject _ascii;")
printer.write("uint8_t _data[2];")
with printer.block('struct', ' latin1[128];'):
printer.write("PyCompactUnicodeObject _latin1;")
printer.write("uint8_t _data[2];")
printer.write(END)
printer.write(after)
def generate_runtime_init(identifiers, strings):
# First get some info from the declarations.
nsmallposints = None
nsmallnegints = None
with open(os.path.join(INTERNAL, 'pycore_global_objects.h')) as infile:
for line in infile:
if line.startswith('#define _PY_NSMALLPOSINTS'):
nsmallposints = int(line.split()[-1])
elif line.startswith('#define _PY_NSMALLNEGINTS'):
nsmallnegints = int(line.split()[-1])
break
else:
raise NotImplementedError
assert nsmallposints and nsmallnegints
# Then target the runtime initializer.
filename = os.path.join(INTERNAL, 'pycore_runtime_init_generated.h')
# Read the non-generated part of the file.
with open(filename) as infile:
orig = infile.read()
lines = iter(orig.rstrip().splitlines())
before = '\n'.join(iter_to_marker(lines, START))
for _ in iter_to_marker(lines, END):
pass
after = '\n'.join(lines)
# Generate the file.
with open_for_changes(filename, orig) as outfile:
immortal_objects = []
printer = Printer(outfile)
printer.write(before)
printer.write(START)
with printer.block('#define _Py_small_ints_INIT', continuation=True):
for i in range(-nsmallnegints, nsmallposints):
printer.write(f'_PyLong_DIGIT_INIT({i}),')
immortal_objects.append(f'(PyObject *)&_Py_SINGLETON(small_ints)[_PY_NSMALLNEGINTS + {i}]')
printer.write('')
with printer.block('#define _Py_bytes_characters_INIT', continuation=True):
for i in range(256):
printer.write(f'_PyBytes_CHAR_INIT({i}),')
immortal_objects.append(f'(PyObject *)&_Py_SINGLETON(bytes_characters)[{i}]')
printer.write('')
with printer.block('#define _Py_str_literals_INIT', continuation=True):
for literal, name in sorted(strings.items(), key=lambda x: x[1]):
printer.write(f'INIT_STR({name}, "{literal}"),')
immortal_objects.append(f'(PyObject *)&_Py_STR({name})')
printer.write('')
with printer.block('#define _Py_str_identifiers_INIT', continuation=True):
for name in sorted(identifiers):
assert name.isidentifier(), name
printer.write(f'INIT_ID({name}),')
immortal_objects.append(f'(PyObject *)&_Py_ID({name})')
printer.write('')
with printer.block('#define _Py_str_ascii_INIT', continuation=True):
for i in range(128):
printer.write(f'_PyASCIIObject_INIT("\\x{i:02x}"),')
immortal_objects.append(f'(PyObject *)&_Py_SINGLETON(strings).ascii[{i}]')
printer.write('')
with printer.block('#define _Py_str_latin1_INIT', continuation=True):
for i in range(128, 256):
utf8 = ['"']
for c in chr(i).encode('utf-8'):
utf8.append(f"\\x{c:02x}")
utf8.append('"')
printer.write(f'_PyUnicode_LATIN1_INIT("\\x{i:02x}", {"".join(utf8)}),')
immortal_objects.append(f'(PyObject *)&_Py_SINGLETON(strings).latin1[{i} - 128]')
printer.write(END)
printer.write(after)
return immortal_objects
def generate_static_strings_initializer(identifiers, strings):
# Target the runtime initializer.
filename = os.path.join(INTERNAL, 'pycore_unicodeobject_generated.h')
# Read the non-generated part of the file.
with open(filename) as infile:
orig = infile.read()
lines = iter(orig.rstrip().splitlines())
before = '\n'.join(iter_to_marker(lines, START))
for _ in iter_to_marker(lines, END):
pass
after = '\n'.join(lines)
# Generate the file.
with open_for_changes(filename, orig) as outfile:
printer = Printer(outfile)
printer.write(before)
printer.write(START)
printer.write("static inline void")
with printer.block("_PyUnicode_InitStaticStrings(PyInterpreterState *interp)"):
printer.write(f'PyObject *string;')
for i in sorted(identifiers):
# This use of _Py_ID() is ignored by iter_global_strings()
# since iter_files() ignores .h files.
printer.write(f'string = &_Py_ID({i});')
printer.write(f'_PyUnicode_InternStatic(interp, &string);')
printer.write(f'assert(_PyUnicode_CheckConsistency(string, 1));')
printer.write(f'assert(PyUnicode_GET_LENGTH(string) != 1);')
for value, name in sorted(strings.items()):
printer.write(f'string = &_Py_STR({name});')
printer.write(f'_PyUnicode_InternStatic(interp, &string);')
printer.write(f'assert(_PyUnicode_CheckConsistency(string, 1));')
printer.write(f'assert(PyUnicode_GET_LENGTH(string) != 1);')
printer.write(END)
printer.write(after)
def generate_global_object_finalizers(generated_immortal_objects):
# Target the runtime initializer.
filename = os.path.join(INTERNAL, 'pycore_global_objects_fini_generated.h')
# Read the non-generated part of the file.
with open(filename) as infile:
orig = infile.read()
lines = iter(orig.rstrip().splitlines())
before = '\n'.join(iter_to_marker(lines, START))
for _ in iter_to_marker(lines, END):
pass
after = '\n'.join(lines)
# Generate the file.
with open_for_changes(filename, orig) as outfile:
printer = Printer(outfile)
printer.write(before)
printer.write(START)
printer.write('#ifdef Py_DEBUG')
printer.write("static inline void")
with printer.block(
"_PyStaticObjects_CheckRefcnt(PyInterpreterState *interp)"):
printer.write('/* generated runtime-global */')
printer.write('// (see pycore_runtime_init_generated.h)')
for ref in generated_immortal_objects:
printer.write(f'_PyStaticObject_CheckRefcnt({ref});')
printer.write('/* non-generated */')
for ref in NON_GENERATED_IMMORTAL_OBJECTS:
printer.write(f'_PyStaticObject_CheckRefcnt({ref});')
printer.write('#endif // Py_DEBUG')
printer.write(END)
printer.write(after)
def get_identifiers_and_strings() -> 'tuple[set[str], dict[str, str]]':
identifiers = set(IDENTIFIERS)
strings = {}
# Note that we store strings as they appear in C source, so the checks here
# can be defeated, e.g.:
# - "a" and "\0x61" won't be reported as duplicate.
# - "\n" appears as 2 characters.
# Probably not worth adding a C string parser.
for name, string, *_ in iter_global_strings():
if string is None:
if name not in IGNORED:
identifiers.add(name)
else:
if len(string) == 1 and ord(string) < 256:
# Give a nice message for common mistakes.
# To cover tricky cases (like "\n") we also generate C asserts.
raise ValueError(
'do not use &_PyID or &_Py_STR for one-character latin-1 '
+ f'strings, use _Py_LATIN1_CHR instead: {string!r}')
if string not in strings:
strings[string] = name
elif name != strings[string]:
raise ValueError(f'string mismatch for {name!r} ({string!r} != {strings[name]!r}')
overlap = identifiers & set(strings.keys())
if overlap:
raise ValueError(
'do not use both _PyID and _Py_DECLARE_STR for the same string: '
+ repr(overlap))
return identifiers, strings
#######################################
# the script
def main() -> None:
identifiers, strings = get_identifiers_and_strings()
generate_global_strings(identifiers, strings)
generated_immortal_objects = generate_runtime_init(identifiers, strings)
generate_static_strings_initializer(identifiers, strings)
generate_global_object_finalizers(generated_immortal_objects)
if __name__ == '__main__':
main()