test_misc.py

# Run the _testcapi module tests (tests for the Python/C API):  by defn,
# these are all functions _testcapi exports whose name begins with 'test_'.

import _thread
from collections import deque
import contextlib
import importlib.machinery
import importlib.util
import json
import os
import pickle
import queue
import random
import sys
import textwrap
import threading
import time
import types
import unittest
import weakref
import operator
from test import support
from test.support import MISSING_C_DOCSTRINGS
from test.support import import_helper
from test.support import script_helper
from test.support import threading_helper
from test.support import warnings_helper
from test.support import requires_limited_api
from test.support import expected_failure_if_gil_disabled
from test.support import Py_GIL_DISABLED
from test.support.script_helper import assert_python_failure, assert_python_ok, run_python_until_end
try:
    import _posixsubprocess
except ImportError:
    _posixsubprocess = None
try:
    import _testmultiphase
except ImportError:
    _testmultiphase = None
try:
    import _testsinglephase
except ImportError:
    _testsinglephase = None
try:
    import _interpreters
except ModuleNotFoundError:
    _interpreters = None

# Skip this test if the _testcapi module isn't available.
_testcapi = import_helper.import_module('_testcapi')

from _testcapi import HeapCTypeSubclass, HeapCTypeSubclassWithFinalizer

import _testlimitedcapi
import _testinternalcapi


NULL = None

def decode_stderr(err):
    return err.decode('utf-8', 'replace').replace('\r', '')


def requires_subinterpreters(meth):
    """Decorator to skip a test if subinterpreters are not supported."""
    return unittest.skipIf(_interpreters is None,
                           'subinterpreters required')(meth)


def testfunction(self):
    """some doc"""
    return self


class InstanceMethod:
    id = _testcapi.instancemethod(id)
    testfunction = _testcapi.instancemethod(testfunction)


CURRENT_THREAD_REGEX = r'Current thread.*:\n' if not support.Py_GIL_DISABLED else r'Stack .*:\n'


@support.force_not_colorized_test_class
class CAPITest(unittest.TestCase):

    def test_instancemethod(self):
        inst = InstanceMethod()
        self.assertEqual(id(inst), inst.id())
        self.assertTrue(inst.testfunction() is inst)
        self.assertEqual(inst.testfunction.__doc__, testfunction.__doc__)
        self.assertEqual(InstanceMethod.testfunction.__doc__, testfunction.__doc__)

        InstanceMethod.testfunction.attribute = "test"
        self.assertEqual(testfunction.attribute, "test")
        self.assertRaises(AttributeError, setattr, inst.testfunction, "attribute", "test")

    @support.requires_subprocess()
    def test_no_FatalError_infinite_loop(self):
        code = textwrap.dedent("""
            import _testcapi
            from test import support

            with support.SuppressCrashReport():
                _testcapi.crash_no_current_thread()
        """)

        run_result, _cmd_line = run_python_until_end('-c', code)
        _rc, out, err = run_result
        self.assertEqual(out, b'')
        # This used to cause an infinite loop.
        if not support.Py_GIL_DISABLED:
            msg = ("Fatal Python error: PyThreadState_Get: "
                   "the function must be called with the GIL held, "
                   "after Python initialization and before Python finalization, "
                   "but the GIL is released "
                   "(the current Python thread state is NULL)").encode()
        else:
            msg = ("Fatal Python error: PyThreadState_Get: "
                   "the function must be called with an active thread state, "
                   "after Python initialization and before Python finalization, "
                   "but it was called without an active thread state. "
                   "Are you trying to call the C API inside of a Py_BEGIN_ALLOW_THREADS block?").encode()
        self.assertStartsWith(err.rstrip(), msg)

    def test_memoryview_from_NULL_pointer(self):
        self.assertRaises(ValueError, _testcapi.make_memoryview_from_NULL_pointer)

    @unittest.skipUnless(_posixsubprocess, '_posixsubprocess required for this test.')
    def test_seq_bytes_to_charp_array(self):
        # Issue #15732: crash in _PySequence_BytesToCharpArray()
        class Z(object):
            def __len__(self):
                return 1
        with self.assertRaisesRegex(TypeError, 'indexing'):
            _posixsubprocess.fork_exec(
                          1,Z(),True,(1, 2),5,6,7,8,9,10,11,12,13,14,True,True,17,False,19,20,21,22)
        # Issue #15736: overflow in _PySequence_BytesToCharpArray()
        class Z(object):
            def __len__(self):
                return sys.maxsize
            def __getitem__(self, i):
                return b'x'
        self.assertRaises(MemoryError, _posixsubprocess.fork_exec,
                          1,Z(),True,(1, 2),5,6,7,8,9,10,11,12,13,14,True,True,17,False,19,20,21,22)

    @unittest.skipUnless(_posixsubprocess, '_posixsubprocess required for this test.')
    def test_subprocess_fork_exec(self):
        class Z(object):
            def __len__(self):
                return 1

        # Issue #15738: crash in subprocess_fork_exec()
        self.assertRaises(TypeError, _posixsubprocess.fork_exec,
                          Z(),[b'1'],True,(1, 2),5,6,7,8,9,10,11,12,13,14,True,True,17,False,19,20,21,22)

    @unittest.skipIf(MISSING_C_DOCSTRINGS,
                     "Signature information for builtins requires docstrings")
    def test_docstring_signature_parsing(self):

        self.assertEqual(_testcapi.no_docstring.__doc__, None)
        self.assertEqual(_testcapi.no_docstring.__text_signature__, None)

        self.assertEqual(_testcapi.docstring_empty.__doc__, None)
        self.assertEqual(_testcapi.docstring_empty.__text_signature__, None)

        self.assertEqual(_testcapi.docstring_no_signature.__doc__,
            "This docstring has no signature.")
        self.assertEqual(_testcapi.docstring_no_signature.__text_signature__, None)

        self.assertEqual(_testcapi.docstring_with_invalid_signature.__doc__,
            "docstring_with_invalid_signature($module, /, boo)\n"
            "\n"
            "This docstring has an invalid signature."
            )
        self.assertEqual(_testcapi.docstring_with_invalid_signature.__text_signature__, None)

        self.assertEqual(_testcapi.docstring_with_invalid_signature2.__doc__,
            "docstring_with_invalid_signature2($module, /, boo)\n"
            "\n"
            "--\n"
            "\n"
            "This docstring also has an invalid signature."
            )
        self.assertEqual(_testcapi.docstring_with_invalid_signature2.__text_signature__, None)

        self.assertEqual(_testcapi.docstring_with_signature.__doc__,
            "This docstring has a valid signature.")
        self.assertEqual(_testcapi.docstring_with_signature.__text_signature__, "($module, /, sig)")

        self.assertEqual(_testcapi.docstring_with_signature_but_no_doc.__doc__, None)
        self.assertEqual(_testcapi.docstring_with_signature_but_no_doc.__text_signature__,
            "($module, /, sig)")

        self.assertEqual(_testcapi.docstring_with_signature_and_extra_newlines.__doc__,
            "\nThis docstring has a valid signature and some extra newlines.")
        self.assertEqual(_testcapi.docstring_with_signature_and_extra_newlines.__text_signature__,
            "($module, /, parameter)")

    def test_c_type_with_matrix_multiplication(self):
        M = _testcapi.matmulType
        m1 = M()
        m2 = M()
        self.assertEqual(m1 @ m2, ("matmul", m1, m2))
        self.assertEqual(m1 @ 42, ("matmul", m1, 42))
        self.assertEqual(42 @ m1, ("matmul", 42, m1))
        o = m1
        o @= m2
        self.assertEqual(o, ("imatmul", m1, m2))
        o = m1
        o @= 42
        self.assertEqual(o, ("imatmul", m1, 42))
        o = 42
        o @= m1
        self.assertEqual(o, ("matmul", 42, m1))

    def test_c_type_with_ipow(self):
        # When the __ipow__ method of a type was implemented in C, using the
        # modulo param would cause segfaults.
        o = _testcapi.ipowType()
        self.assertEqual(o.__ipow__(1), (1, None))
        self.assertEqual(o.__ipow__(2, 2), (2, 2))

    def test_return_null_without_error(self):
        # Issue #23571: A function must not return NULL without setting an
        # error
        if support.Py_DEBUG:
            code = textwrap.dedent("""
                import _testcapi
                from test import support

                with support.SuppressCrashReport():
                    _testcapi.return_null_without_error()
            """)
            rc, out, err = assert_python_failure('-c', code)
            err = decode_stderr(err)
            self.assertRegex(err,
                r'Fatal Python error: _Py_CheckFunctionResult: '
                    r'a function returned NULL without setting an exception\n'
                r'Python runtime state: initialized\n'
                r'SystemError: <built-in function return_null_without_error> '
                    r'returned NULL without setting an exception\n'
                r'\n' +
                CURRENT_THREAD_REGEX +
                r'  File .*", line 6 in <module>\n')
        else:
            with self.assertRaises(SystemError) as cm:
                _testcapi.return_null_without_error()
            self.assertRegex(str(cm.exception),
                             'return_null_without_error.* '
                             'returned NULL without setting an exception')

    def test_return_result_with_error(self):
        # Issue #23571: A function must not return a result with an error set
        if support.Py_DEBUG:
            code = textwrap.dedent("""
                import _testcapi
                from test import support

                with support.SuppressCrashReport():
                    _testcapi.return_result_with_error()
            """)
            rc, out, err = assert_python_failure('-c', code)
            err = decode_stderr(err)
            self.assertRegex(err,
                    r'Fatal Python error: _Py_CheckFunctionResult: '
                        r'a function returned a result with an exception set\n'
                    r'Python runtime state: initialized\n'
                    r'ValueError\n'
                    r'\n'
                    r'The above exception was the direct cause '
                        r'of the following exception:\n'
                    r'\n'
                    r'SystemError: <built-in '
                        r'function return_result_with_error> '
                        r'returned a result with an exception set\n'
                    r'\n' +
                    CURRENT_THREAD_REGEX +
                    r'  File .*, line 6 in <module>\n')
        else:
            with self.assertRaises(SystemError) as cm:
                _testcapi.return_result_with_error()
            self.assertRegex(str(cm.exception),
                             'return_result_with_error.* '
                             'returned a result with an exception set')

    def test_getitem_with_error(self):
        # Test _Py_CheckSlotResult(). Raise an exception and then calls
        # PyObject_GetItem(): check that the assertion catches the bug.
        # PyObject_GetItem() must not be called with an exception set.
        code = textwrap.dedent("""
            import _testcapi
            from test import support

            with support.SuppressCrashReport():
                _testcapi.getitem_with_error({1: 2}, 1)
        """)
        rc, out, err = assert_python_failure('-c', code)
        err = decode_stderr(err)
        if 'SystemError: ' not in err:
            self.assertRegex(err,
                    r'Fatal Python error: _Py_CheckSlotResult: '
                        r'Slot __getitem__ of type dict succeeded '
                        r'with an exception set\n'
                    r'Python runtime state: initialized\n'
                    r'ValueError: bug\n'
                    r'\n' +
                    CURRENT_THREAD_REGEX +
                    r'  File .*, line 6 in <module>\n'
                    r'\n'
                    r'Extension modules: ')
        else:
            # Python built with NDEBUG macro defined:
            # test _Py_CheckFunctionResult() instead.
            self.assertIn('returned a result with an exception set', err)

    def test_buildvalue(self):
        # Test Py_BuildValue() with object arguments
        buildvalue = _testcapi.py_buildvalue
        self.assertEqual(buildvalue(''), None)
        self.assertEqual(buildvalue('()'), ())
        self.assertEqual(buildvalue('[]'), [])
        self.assertEqual(buildvalue('{}'), {})
        self.assertEqual(buildvalue('()[]{}'), ((), [], {}))
        self.assertEqual(buildvalue('O', 1), 1)
        self.assertEqual(buildvalue('(O)', 1), (1,))
        self.assertEqual(buildvalue('[O]', 1), [1])
        self.assertRaises(SystemError, buildvalue, '{O}', 1)
        self.assertEqual(buildvalue('OO', 1, 2), (1, 2))
        self.assertEqual(buildvalue('(OO)', 1, 2), (1, 2))
        self.assertEqual(buildvalue('[OO]', 1, 2), [1, 2])
        self.assertEqual(buildvalue('{OO}', 1, 2), {1: 2})
        self.assertEqual(buildvalue('{OOOO}', 1, 2, 3, 4), {1: 2, 3: 4})
        self.assertEqual(buildvalue('((O))', 1), ((1,),))
        self.assertEqual(buildvalue('((OO))', 1, 2), ((1, 2),))

        self.assertEqual(buildvalue(' \t,:'), None)
        self.assertEqual(buildvalue('O,', 1), 1)
        self.assertEqual(buildvalue('   O   ', 1), 1)
        self.assertEqual(buildvalue('\tO\t', 1), 1)
        self.assertEqual(buildvalue('O,O', 1, 2), (1, 2))
        self.assertEqual(buildvalue('O, O', 1, 2), (1, 2))
        self.assertEqual(buildvalue('O,\tO', 1, 2), (1, 2))
        self.assertEqual(buildvalue('O O', 1, 2), (1, 2))
        self.assertEqual(buildvalue('O\tO', 1, 2), (1, 2))
        self.assertEqual(buildvalue('(O,O)', 1, 2), (1, 2))
        self.assertEqual(buildvalue('(O, O,)', 1, 2), (1, 2))
        self.assertEqual(buildvalue(' ( O O ) ', 1, 2), (1, 2))
        self.assertEqual(buildvalue('\t(\tO\tO\t)\t', 1, 2), (1, 2))
        self.assertEqual(buildvalue('[O,O]', 1, 2), [1, 2])
        self.assertEqual(buildvalue('[O, O,]', 1, 2), [1, 2])
        self.assertEqual(buildvalue(' [ O O ] ', 1, 2), [1, 2])
        self.assertEqual(buildvalue(' [\tO\tO\t] ', 1, 2), [1, 2])
        self.assertEqual(buildvalue('{O:O}', 1, 2), {1: 2})
        self.assertEqual(buildvalue('{O:O,O:O}', 1, 2, 3, 4), {1: 2, 3: 4})
        self.assertEqual(buildvalue('{O: O, O: O,}', 1, 2, 3, 4), {1: 2, 3: 4})
        self.assertEqual(buildvalue(' { O O O O } ', 1, 2, 3, 4), {1: 2, 3: 4})
        self.assertEqual(buildvalue('\t{\tO\tO\tO\tO\t}\t', 1, 2, 3, 4), {1: 2, 3: 4})

        self.assertRaises(SystemError, buildvalue, 'O', NULL)
        self.assertRaises(SystemError, buildvalue, '(O)', NULL)
        self.assertRaises(SystemError, buildvalue, '[O]', NULL)
        self.assertRaises(SystemError, buildvalue, '{O}', NULL)
        self.assertRaises(SystemError, buildvalue, 'OO', 1, NULL)
        self.assertRaises(SystemError, buildvalue, 'OO', NULL, 2)
        self.assertRaises(SystemError, buildvalue, '(OO)', 1, NULL)
        self.assertRaises(SystemError, buildvalue, '(OO)', NULL, 2)
        self.assertRaises(SystemError, buildvalue, '[OO]', 1, NULL)
        self.assertRaises(SystemError, buildvalue, '[OO]', NULL, 2)
        self.assertRaises(SystemError, buildvalue, '{OO}', 1, NULL)
        self.assertRaises(SystemError, buildvalue, '{OO}', NULL, 2)

    def test_buildvalue_ints(self):
        # Test Py_BuildValue() with integer arguments
        buildvalue = _testcapi.py_buildvalue_ints
        from _testcapi import SHRT_MIN, SHRT_MAX, USHRT_MAX, INT_MIN, INT_MAX, UINT_MAX
        self.assertEqual(buildvalue('i', INT_MAX), INT_MAX)
        self.assertEqual(buildvalue('i', INT_MIN), INT_MIN)
        self.assertEqual(buildvalue('I', UINT_MAX), UINT_MAX)

        self.assertEqual(buildvalue('h', SHRT_MAX), SHRT_MAX)
        self.assertEqual(buildvalue('h', SHRT_MIN), SHRT_MIN)
        self.assertEqual(buildvalue('H', USHRT_MAX), USHRT_MAX)

        self.assertEqual(buildvalue('b', 127), 127)
        self.assertEqual(buildvalue('b', -128), -128)
        self.assertEqual(buildvalue('B', 255), 255)

        self.assertEqual(buildvalue('c', ord('A')), b'A')
        self.assertEqual(buildvalue('c', 255), b'\xff')
        self.assertEqual(buildvalue('c', 256), b'\x00')
        self.assertEqual(buildvalue('c', -1), b'\xff')

        self.assertEqual(buildvalue('C', 255), chr(255))
        self.assertEqual(buildvalue('C', 256), chr(256))
        self.assertEqual(buildvalue('C', sys.maxunicode), chr(sys.maxunicode))
        self.assertRaises(ValueError, buildvalue, 'C', -1)
        self.assertRaises(ValueError, buildvalue, 'C', sys.maxunicode+1)

        # gh-84489
        self.assertRaises(ValueError, buildvalue, '(C )i', -1, 2)
        self.assertRaises(ValueError, buildvalue, '[C ]i', -1, 2)
        self.assertRaises(ValueError, buildvalue, '{Ci }i', -1, 2, 3)

    def test_buildvalue_N(self):
        _testcapi.test_buildvalue_N()

    def test_trashcan_subclass(self):
        # bpo-35983: Check that the trashcan mechanism for "list" is NOT
        # activated when its tp_dealloc is being called by a subclass
        from _testcapi import MyList
        L = None
        for i in range(100):
            L = MyList((L,))

    @support.requires_resource('cpu')
    @support.skip_emscripten_stack_overflow()
    @support.skip_wasi_stack_overflow()
    def test_trashcan_python_class1(self):
        self.do_test_trashcan_python_class(list)

    @support.requires_resource('cpu')
    @support.skip_emscripten_stack_overflow()
    @support.skip_wasi_stack_overflow()
    def test_trashcan_python_class2(self):
        from _testcapi import MyList
        self.do_test_trashcan_python_class(MyList)

    def do_test_trashcan_python_class(self, base):
        # Check that the trashcan mechanism works properly for a Python
        # subclass of a class using the trashcan (this specific test assumes
        # that the base class "base" behaves like list)
        class PyList(base):
            # Count the number of PyList instances to verify that there is
            # no memory leak
            num = 0
            def __init__(self, *args):
                __class__.num += 1
                super().__init__(*args)
            def __del__(self):
                __class__.num -= 1

        for parity in (0, 1):
            L = None
            # We need in the order of 2**20 iterations here such that a
            # typical 8MB stack would overflow without the trashcan.
            for i in range(2**20):
                L = PyList((L,))
                L.attr = i
            if parity:
                # Add one additional nesting layer
                L = (L,)
            self.assertGreater(PyList.num, 0)
            del L
            self.assertEqual(PyList.num, 0)

    @unittest.skipIf(MISSING_C_DOCSTRINGS,
                     "Signature information for builtins requires docstrings")
    def test_heap_ctype_doc_and_text_signature(self):
        self.assertEqual(_testcapi.HeapDocCType.__doc__, "somedoc")
        self.assertEqual(_testcapi.HeapDocCType.__text_signature__, "(arg1, arg2)")

    def test_null_type_doc(self):
        self.assertEqual(_testcapi.NullTpDocType.__doc__, None)

    def test_subclass_of_heap_gc_ctype_with_tpdealloc_decrefs_once(self):
        class HeapGcCTypeSubclass(_testcapi.HeapGcCType):
            def __init__(self):
                self.value2 = 20
                super().__init__()

        subclass_instance = HeapGcCTypeSubclass()
        type_refcnt = sys.getrefcount(HeapGcCTypeSubclass)

        # Test that subclass instance was fully created
        self.assertEqual(subclass_instance.value, 10)
        self.assertEqual(subclass_instance.value2, 20)

        # Test that the type reference count is only decremented once
        del subclass_instance
        self.assertEqual(type_refcnt - 1, sys.getrefcount(HeapGcCTypeSubclass))

    def test_subclass_of_heap_gc_ctype_with_del_modifying_dunder_class_only_decrefs_once(self):
        class A(_testcapi.HeapGcCType):
            def __init__(self):
                self.value2 = 20
                super().__init__()

        class B(A):
            def __init__(self):
                super().__init__()

            def __del__(self):
                self.__class__ = A
                A.refcnt_in_del = sys.getrefcount(A)
                B.refcnt_in_del = sys.getrefcount(B)

        subclass_instance = B()
        type_refcnt = sys.getrefcount(B)
        new_type_refcnt = sys.getrefcount(A)

        # Test that subclass instance was fully created
        self.assertEqual(subclass_instance.value, 10)
        self.assertEqual(subclass_instance.value2, 20)

        del subclass_instance

        # Test that setting __class__ modified the reference counts of the types
        if support.Py_DEBUG:
            # gh-89373: In debug mode, _Py_Dealloc() keeps a strong reference
            # to the type while calling tp_dealloc()
            self.assertEqual(type_refcnt, B.refcnt_in_del)
        else:
            self.assertEqual(type_refcnt - 1, B.refcnt_in_del)
        self.assertEqual(new_type_refcnt + 1, A.refcnt_in_del)

        # Test that the original type already has decreased its refcnt
        self.assertEqual(type_refcnt - 1, sys.getrefcount(B))

        # Test that subtype_dealloc decref the newly assigned __class__ only once
        self.assertEqual(new_type_refcnt, sys.getrefcount(A))

    def test_heaptype_with_dict(self):
        for cls in (
            _testcapi.HeapCTypeWithDict,
            _testlimitedcapi.HeapCTypeWithRelativeDict,
        ):
            with self.subTest(cls=cls):
                inst = cls()
                inst.foo = 42
                self.assertEqual(inst.foo, 42)
                self.assertEqual(inst.dictobj, inst.__dict__)
                self.assertEqual(inst.dictobj, {"foo": 42})

                inst = cls()
                self.assertEqual({}, inst.__dict__)

    def test_heaptype_with_managed_dict(self):
        inst = _testcapi.HeapCTypeWithManagedDict()
        inst.foo = 42
        self.assertEqual(inst.foo, 42)
        self.assertEqual(inst.__dict__, {"foo": 42})

        inst = _testcapi.HeapCTypeWithManagedDict()
        self.assertEqual({}, inst.__dict__)

        a = _testcapi.HeapCTypeWithManagedDict()
        b = _testcapi.HeapCTypeWithManagedDict()
        a.b = b
        b.a = a
        del a, b

    def test_sublclassing_managed_dict(self):

        class C(_testcapi.HeapCTypeWithManagedDict):
            pass

        i = C()
        i.spam = i
        del i

    def test_heaptype_with_negative_dict(self):
        inst = _testcapi.HeapCTypeWithNegativeDict()
        inst.foo = 42
        self.assertEqual(inst.foo, 42)
        self.assertEqual(inst.dictobj, inst.__dict__)
        self.assertEqual(inst.dictobj, {"foo": 42})

        inst = _testcapi.HeapCTypeWithNegativeDict()
        self.assertEqual({}, inst.__dict__)

    def test_heaptype_with_weakref(self):
        for cls in (
            _testcapi.HeapCTypeWithWeakref,
            _testlimitedcapi.HeapCTypeWithRelativeWeakref,
        ):
            with self.subTest(cls=cls):
                inst = cls()
                ref = weakref.ref(inst)
                self.assertEqual(ref(), inst)
                self.assertEqual(inst.weakreflist, ref)

    def test_heaptype_with_managed_weakref(self):
        inst = _testcapi.HeapCTypeWithManagedWeakref()
        ref = weakref.ref(inst)
        self.assertEqual(ref(), inst)

    def test_sublclassing_managed_weakref(self):

        class C(_testcapi.HeapCTypeWithManagedWeakref):
            pass

        inst = C()
        ref = weakref.ref(inst)
        self.assertEqual(ref(), inst)

    def test_sublclassing_managed_both(self):

        class C1(_testcapi.HeapCTypeWithManagedWeakref, _testcapi.HeapCTypeWithManagedDict):
            pass

        class C2(_testcapi.HeapCTypeWithManagedDict, _testcapi.HeapCTypeWithManagedWeakref):
            pass

        for cls in (C1, C2):
            inst = cls()
            ref = weakref.ref(inst)
            self.assertEqual(ref(), inst)
            inst.spam = inst
            del inst
            ref = weakref.ref(cls())
            self.assertIs(ref(), None)

    def test_heaptype_with_buffer(self):
        inst = _testcapi.HeapCTypeWithBuffer()
        b = bytes(inst)
        self.assertEqual(b, b"1234")

    def test_c_subclass_of_heap_ctype_with_tpdealloc_decrefs_once(self):
        subclass_instance = _testcapi.HeapCTypeSubclass()
        type_refcnt = sys.getrefcount(_testcapi.HeapCTypeSubclass)

        # Test that subclass instance was fully created
        self.assertEqual(subclass_instance.value, 10)
        self.assertEqual(subclass_instance.value2, 20)

        # Test that the type reference count is only decremented once
        del subclass_instance
        self.assertEqual(type_refcnt - 1, sys.getrefcount(_testcapi.HeapCTypeSubclass))

    def test_c_subclass_of_heap_ctype_with_del_modifying_dunder_class_only_decrefs_once(self):
        subclass_instance = HeapCTypeSubclassWithFinalizer()
        type_refcnt = sys.getrefcount(HeapCTypeSubclassWithFinalizer)
        new_type_refcnt = sys.getrefcount(HeapCTypeSubclass)

        # Test that subclass instance was fully created
        self.assertEqual(subclass_instance.value, 10)
        self.assertEqual(subclass_instance.value2, 20)

        # The tp_finalize slot will set __class__ to HeapCTypeSubclass
        del subclass_instance

        # Test that setting __class__ modified the reference counts of the types
        #
        # This is highly sensitive to implementation details and may break in the future.
        #
        # We expect the refcount on the old type, HeapCTypeSubclassWithFinalizer, to
        # remain the same: the finalizer gets a strong reference (+1) when it gets the
        # type from the module and setting __class__ decrements the refcount (-1).
        #
        # We expect the refcount on the new type, HeapCTypeSubclass, to increase by 2:
        # the finalizer get a strong reference (+1) when it gets the type from the
        # module and setting __class__ increments the refcount (+1).
        expected_type_refcnt = type_refcnt
        expected_new_type_refcnt = new_type_refcnt + 2

        if not Py_GIL_DISABLED:
            # In default builds the result returned from sys.getrefcount
            # includes a temporary reference that is created by the interpreter
            # when it pushes its argument on the operand stack. This temporary
            # reference is not included in the result returned by Py_REFCNT, which
            # is used in the finalizer.
            #
            # In free-threaded builds the result returned from sys.getrefcount
            # does not include the temporary reference. Types use deferred
            # refcounting and the interpreter will not create a new reference
            # for deferred values on the operand stack.
            expected_type_refcnt -= 1
            expected_new_type_refcnt -= 1

        if support.Py_DEBUG:
            # gh-89373: In debug mode, _Py_Dealloc() keeps a strong reference
            # to the type while calling tp_dealloc()
            expected_type_refcnt += 1

        self.assertEqual(expected_type_refcnt, HeapCTypeSubclassWithFinalizer.refcnt_in_del)
        self.assertEqual(expected_new_type_refcnt, HeapCTypeSubclass.refcnt_in_del)

        # Test that the original type already has decreased its refcnt
        self.assertEqual(type_refcnt - 1, sys.getrefcount(HeapCTypeSubclassWithFinalizer))

        # Test that subtype_dealloc decref the newly assigned __class__ only once
        self.assertEqual(new_type_refcnt, sys.getrefcount(HeapCTypeSubclass))

    def test_heaptype_with_setattro(self):
        obj = _testcapi.HeapCTypeSetattr()
        self.assertEqual(obj.pvalue, 10)
        obj.value = 12
        self.assertEqual(obj.pvalue, 12)
        del obj.value
        self.assertEqual(obj.pvalue, 0)

    def test_heaptype_with_custom_metaclass(self):
        metaclass = _testcapi.HeapCTypeMetaclass
        self.assertIsSubclass(metaclass, type)

        # Class creation from C
        t = _testcapi.pytype_fromspec_meta(metaclass)
        self.assertIsInstance(t, type)
        self.assertEqual(t.__name__, "HeapCTypeViaMetaclass")
        self.assertIs(type(t), metaclass)

        # Class creation from Python
        t = metaclass("PyClassViaMetaclass", (), {})
        self.assertIsInstance(t, type)
        self.assertEqual(t.__name__, "PyClassViaMetaclass")

    def test_heaptype_with_custom_metaclass_null_new(self):
        metaclass = _testcapi.HeapCTypeMetaclassNullNew

        self.assertIsSubclass(metaclass, type)

        # Class creation from C
        t = _testcapi.pytype_fromspec_meta(metaclass)
        self.assertIsInstance(t, type)
        self.assertEqual(t.__name__, "HeapCTypeViaMetaclass")
        self.assertIs(type(t), metaclass)

        # Class creation from Python
        with self.assertRaisesRegex(TypeError, "cannot create .* instances"):
            metaclass("PyClassViaMetaclass", (), {})

    def test_heaptype_with_custom_metaclass_custom_new(self):
        metaclass = _testcapi.HeapCTypeMetaclassCustomNew

        self.assertIsSubclass(_testcapi.HeapCTypeMetaclassCustomNew, type)

        msg = "Metaclasses with custom tp_new are not supported."
        with self.assertRaisesRegex(TypeError, msg):
            t = _testcapi.pytype_fromspec_meta(metaclass)

    def test_heaptype_base_with_custom_metaclass(self):
        metaclass = _testcapi.HeapCTypeMetaclassCustomNew

        class Base(metaclass=metaclass):
            pass

        # Class creation from C
        msg = "Metaclasses with custom tp_new are not supported."
        with self.assertRaisesRegex(TypeError, msg):
            sub = _testcapi.make_type_with_base(Base)

    def test_heaptype_with_tp_vectorcall(self):
        tp = _testcapi.HeapCTypeVectorcall
        v0 = tp.__new__(tp)
        v0.__init__()
        v1 = tp()
        self.assertEqual(v0.value, 2)
        self.assertEqual(v1.value, 1)

    def test_multiple_inheritance_ctypes_with_weakref_or_dict(self):
        for weakref_cls in (_testcapi.HeapCTypeWithWeakref,
                            _testlimitedcapi.HeapCTypeWithRelativeWeakref):
            for dict_cls in (_testcapi.HeapCTypeWithDict,
                             _testlimitedcapi.HeapCTypeWithRelativeDict):
                with self.subTest(weakref_cls=weakref_cls, dict_cls=dict_cls):

                    with self.assertRaises(TypeError):
                        class Both1(weakref_cls, dict_cls):
                            pass
                    with self.assertRaises(TypeError):
                        class Both2(dict_cls, weakref_cls):
                            pass

    def test_multiple_inheritance_ctypes_with_weakref_or_dict_and_other_builtin(self):
        for dict_cls in (_testcapi.HeapCTypeWithDict,
                         _testlimitedcapi.HeapCTypeWithRelativeDict):
            for weakref_cls in (_testcapi.HeapCTypeWithWeakref,
                                _testlimitedcapi.HeapCTypeWithRelativeWeakref):
                with self.subTest(dict_cls=dict_cls, weakref_cls=weakref_cls):

                    with self.assertRaises(TypeError):
                        class C1(dict_cls, list):
                            pass

                    with self.assertRaises(TypeError):
                        class C2(weakref_cls, list):
                            pass

                    class C3(_testcapi.HeapCTypeWithManagedDict, list):
                        pass
                    class C4(_testcapi.HeapCTypeWithManagedWeakref, list):
                        pass

                    inst = C3()
                    inst.append(0)
                    str(inst.__dict__)

                    inst = C4()
                    inst.append(0)
                    str(inst.__weakref__)

                    for cls in (_testcapi.HeapCTypeWithManagedDict,
                                _testcapi.HeapCTypeWithManagedWeakref):
                        for cls2 in (dict_cls, weakref_cls):
                            class S(cls, cls2):
                                pass
                        class B1(C3, cls):
                            pass
                        class B2(C4, cls):
                            pass

    def test_pytype_fromspec_with_repeated_slots(self):
        for variant in range(2):
            with self.subTest(variant=variant):
                with self.assertRaises(SystemError):
                    _testcapi.create_type_from_repeated_slots(variant)

    def test_immutable_type_with_mutable_base(self):
        class MutableBase: ...

        with self.assertRaisesRegex(TypeError, 'Creating immutable type'):
            _testcapi.make_immutable_type_with_base(MutableBase)

    def test_pynumber_tobase(self):
        from _testcapi import pynumber_tobase
        small_number = 123
        large_number = 2**64
        class IDX:
            def __init__(self, val):
                self.val = val
            def __index__(self):
                return self.val

        test_cases = ((2, '0b1111011', '0b10000000000000000000000000000000000000000000000000000000000000000'),
                      (8, '0o173', '0o2000000000000000000000'),
                      (10, '123', '18446744073709551616'),
                      (16, '0x7b', '0x10000000000000000'))
        for base, small_target, large_target in test_cases:
            with self.subTest(base=base, st=small_target, lt=large_target):
                # Test for small number
                self.assertEqual(pynumber_tobase(small_number, base), small_target)
                self.assertEqual(pynumber_tobase(-small_number, base), '-' + small_target)
                self.assertEqual(pynumber_tobase(IDX(small_number), base), small_target)
                # Test for large number(out of range of a longlong,i.e.[-2**63, 2**63-1])
                self.assertEqual(pynumber_tobase(large_number, base), large_target)
                self.assertEqual(pynumber_tobase(-large_number, base), '-' + large_target)
                self.assertEqual(pynumber_tobase(IDX(large_number), base), large_target)
        self.assertRaises(TypeError, pynumber_tobase, IDX(123.0), 10)
        self.assertRaises(TypeError, pynumber_tobase, IDX('123'), 10)
        self.assertRaises(TypeError, pynumber_tobase, 123.0, 10)
        self.assertRaises(TypeError, pynumber_tobase, '123', 10)
        self.assertRaises(SystemError, pynumber_tobase, 123, 0)

    def test_pyobject_repr_from_null(self):
        s = _testcapi.pyobject_repr_from_null()
        self.assertEqual(s, '<NULL>')

    def test_pyobject_str_from_null(self):
        s = _testcapi.pyobject_str_from_null()
        self.assertEqual(s, '<NULL>')

    def test_pyobject_bytes_from_null(self):
        s = _testcapi.pyobject_bytes_from_null()
        self.assertEqual(s, b'<NULL>')

    def test_Py_CompileString(self):
        # Check that Py_CompileString respects the coding cookie
        _compile = _testcapi.Py_CompileString
        code = b"# -*- coding: latin1 -*-\nprint('\xc2\xa4')\n"
        result = _compile(code)
        expected = compile(code, "<string>", "exec")
        self.assertEqual(result.co_consts, expected.co_consts)

    def test_export_symbols(self):
        # bpo-44133: Ensure that the "Py_FrozenMain" and
        # "PyThread_get_thread_native_id" symbols are exported by the Python
        # (directly by the binary, or via by the Python dynamic library).
        ctypes = import_helper.import_module('ctypes')
        names = []

        # Test if the PY_HAVE_THREAD_NATIVE_ID macro is defined
        if hasattr(_thread, 'get_native_id'):
            names.append('PyThread_get_thread_native_id')

        # Python/frozenmain.c fails to build on Windows when the symbols are
        # missing:
        # - PyWinFreeze_ExeInit
        # - PyWinFreeze_ExeTerm
        # - PyInitFrozenExtensions
        if os.name != 'nt':
            names.append('Py_FrozenMain')

        for name in names:
            self.assertHasAttr(ctypes.pythonapi, name)

    def test_clear_managed_dict(self):

        class C:
            def __init__(self):
                self.a = 1

        c = C()
        _testcapi.clear_managed_dict(c)
        self.assertEqual(c.__dict__, {})
        c = C()
        self.assertEqual(c.__dict__, {'a':1})
        _testcapi.clear_managed_dict(c)
        self.assertEqual(c.__dict__, {})

    def test_unstable_gc_new_with_extra_data(self):
        class Data(_testcapi.ObjExtraData):
            __slots__ = ('x', 'y')

        d = Data()
        d.x = 10
        d.y = 20
        d.extra = 30
        self.assertEqual(d.x, 10)
        self.assertEqual(d.y, 20)
        self.assertEqual(d.extra, 30)
        del d.extra
        self.assertIsNone(d.extra)

    def test_gen_get_code(self):
        def genf(): yield
        gen = genf()
        self.assertEqual(_testcapi.gen_get_code(gen), gen.gi_code)

    def test_tp_bases_slot(self):
        cls = _testcapi.HeapCTypeWithBasesSlot
        self.assertEqual(cls.__bases__, (int,))
        self.assertEqual(cls.__base__, int)

    def test_tp_bases_slot_none(self):
        self.assertRaisesRegex(
            SystemError,
            "Py_tp_bases is not a tuple",
            _testcapi.create_heapctype_with_none_bases_slot
        )


@requires_limited_api
class TestHeapTypeRelative(unittest.TestCase):
    """Test API for extending opaque types (PEP 697)"""

    @requires_limited_api
    def test_heaptype_relative_sizes(self):
        # Test subclassing using "relative" basicsize, see PEP 697
        def check(extra_base_size, extra_size):
            Base, Sub, instance, data_ptr, data_offset, data_size = (
                _testlimitedcapi.make_sized_heaptypes(
                    extra_base_size, -extra_size))

            # no alignment shenanigans when inheriting directly
            if extra_size == 0:
                self.assertEqual(Base.__basicsize__, Sub.__basicsize__)
                self.assertEqual(data_size, 0)

            else:
                # The following offsets should be in increasing order:
                offsets = [
                    (0, 'start of object'),
                    (Base.__basicsize__, 'end of base data'),
                    (data_offset, 'subclass data'),
                    (data_offset + extra_size, 'end of requested subcls data'),
                    (data_offset + data_size, 'end of reserved subcls data'),
                    (Sub.__basicsize__, 'end of object'),
                ]
                ordered_offsets = sorted(offsets, key=operator.itemgetter(0))
                self.assertEqual(
                    offsets, ordered_offsets,
                    msg=f'Offsets not in expected order, got: {ordered_offsets}')

                # end of reserved subcls data == end of object
                self.assertEqual(Sub.__basicsize__, data_offset + data_size)

                # we don't reserve (requested + alignment) or more data
                self.assertLess(data_size - extra_size,
                                _testlimitedcapi.ALIGNOF_MAX_ALIGN_T)

            # The offsets/sizes we calculated should be aligned.
            self.assertEqual(data_offset % _testlimitedcapi.ALIGNOF_MAX_ALIGN_T, 0)
            self.assertEqual(data_size % _testlimitedcapi.ALIGNOF_MAX_ALIGN_T, 0)

        sizes = sorted({0, 1, 2, 3, 4, 7, 8, 123,
                        object.__basicsize__,
                        object.__basicsize__-1,
                        object.__basicsize__+1})
        for extra_base_size in sizes:
            for extra_size in sizes:
                args = dict(extra_base_size=extra_base_size,
                            extra_size=extra_size)
                with self.subTest(**args):
                    check(**args)

    def test_HeapCCollection(self):
        """Make sure HeapCCollection works properly by itself"""
        collection = _testcapi.HeapCCollection(1, 2, 3)
        self.assertEqual(list(collection), [1, 2, 3])

    def test_heaptype_inherit_itemsize(self):
        """Test HeapCCollection subclasses work properly"""
        sizes = sorted({0, 1, 2, 3, 4, 7, 8, 123,
                        object.__basicsize__,
                        object.__basicsize__-1,
                        object.__basicsize__+1})
        for extra_size in sizes:
            with self.subTest(extra_size=extra_size):
                Sub = _testlimitedcapi.subclass_var_heaptype(
                    _testcapi.HeapCCollection, -extra_size, 0, 0)
                collection = Sub(1, 2, 3)
                collection.set_data_to_3s()

                self.assertEqual(list(collection), [1, 2, 3])
                mem = collection.get_data()
                self.assertGreaterEqual(len(mem), extra_size)
                self.assertTrue(set(mem) <= {3}, f'got {mem!r}')

    def test_heaptype_invalid_inheritance(self):
        with self.assertRaises(SystemError,
                               msg="Cannot extend variable-size class without "
                               + "Py_TPFLAGS_ITEMS_AT_END"):
            _testlimitedcapi.subclass_heaptype(int, -8, 0)

    def test_heaptype_relative_members(self):
        """Test HeapCCollection subclasses work properly"""
        sizes = sorted({0, 1, 2, 3, 4, 7, 8, 123,
                        object.__basicsize__,
                        object.__basicsize__-1,
                        object.__basicsize__+1})
        for extra_base_size in sizes:
            for extra_size in sizes:
                for offset in sizes:
                    with self.subTest(extra_base_size=extra_base_size, extra_size=extra_size, offset=offset):
                        if offset < extra_size:
                            Sub = _testlimitedcapi.make_heaptype_with_member(
                                extra_base_size, -extra_size, offset, True)
                            Base = Sub.mro()[1]
                            instance = Sub()
                            self.assertEqual(instance.memb, instance.get_memb())
                            instance.set_memb(13)
                            self.assertEqual(instance.memb, instance.get_memb())
                            self.assertEqual(instance.get_memb(), 13)
                            instance.memb = 14
                            self.assertEqual(instance.memb, instance.get_memb())
                            self.assertEqual(instance.get_memb(), 14)
                            self.assertGreaterEqual(instance.get_memb_offset(), Base.__basicsize__)
                            self.assertLess(instance.get_memb_offset(), Sub.__basicsize__)
                            with self.assertRaises(SystemError):
                                instance.get_memb_relative()
                            with self.assertRaises(SystemError):
                                instance.set_memb_relative(0)
                        else:
                            with self.assertRaises(SystemError):
                                Sub = _testlimitedcapi.make_heaptype_with_member(
                                    extra_base_size, -extra_size, offset, True)
                        with self.assertRaises(SystemError):
                            Sub = _testlimitedcapi.make_heaptype_with_member(
                                extra_base_size, extra_size, offset, True)
                with self.subTest(extra_base_size=extra_base_size, extra_size=extra_size):
                    with self.assertRaises(SystemError):
                        Sub = _testlimitedcapi.make_heaptype_with_member(
                            extra_base_size, -extra_size, -1, True)

    def test_heaptype_relative_members_errors(self):
        with self.assertRaisesRegex(
                SystemError,
                r"With Py_RELATIVE_OFFSET, basicsize must be negative"):
            _testlimitedcapi.make_heaptype_with_member(0, 1234, 0, True)
        with self.assertRaisesRegex(
                SystemError, r"Member offset out of range \(0\.\.-basicsize\)"):
            _testlimitedcapi.make_heaptype_with_member(0, -8, 1234, True)
        with self.assertRaisesRegex(
                SystemError, r"Member offset out of range \(0\.\.-basicsize\)"):
            _testlimitedcapi.make_heaptype_with_member(0, -8, -1, True)

        Sub = _testlimitedcapi.make_heaptype_with_member(0, -8, 0, True)
        instance = Sub()
        with self.assertRaisesRegex(
                SystemError, r"PyMember_GetOne used with Py_RELATIVE_OFFSET"):
            instance.get_memb_relative()
        with self.assertRaisesRegex(
                SystemError, r"PyMember_SetOne used with Py_RELATIVE_OFFSET"):
            instance.set_memb_relative(0)

    def test_heaptype_relative_special_members_errors(self):
        for member_name in "__vectorcalloffset__", "__dictoffset__", "__weaklistoffset__":
            with self.subTest(member_name=member_name):
                with self.assertRaisesRegex(
                        SystemError,
                        r"With Py_RELATIVE_OFFSET, basicsize must be negative."):
                    _testlimitedcapi.make_heaptype_with_member(
                        basicsize=sys.getsizeof(object()) + 100,
                        add_relative_flag=True,
                        member_name=member_name,
                        member_offset=0,
                        member_type=_testlimitedcapi.Py_T_PYSSIZET,
                        member_flags=_testlimitedcapi.Py_READONLY,
                        )
                with self.assertRaisesRegex(
                        SystemError,
                        r"Member offset out of range \(0\.\.-basicsize\)"):
                    _testlimitedcapi.make_heaptype_with_member(
                        basicsize=-8,
                        add_relative_flag=True,
                        member_name=member_name,
                        member_offset=-1,
                        member_type=_testlimitedcapi.Py_T_PYSSIZET,
                        member_flags=_testlimitedcapi.Py_READONLY,
                        )
                with self.assertRaisesRegex(
                        SystemError,
                        r"type of %s must be Py_T_PYSSIZET" % member_name):
                    _testlimitedcapi.make_heaptype_with_member(
                        basicsize=-100,
                        add_relative_flag=True,
                        member_name=member_name,
                        member_offset=0,
                        member_flags=_testlimitedcapi.Py_READONLY,
                        )
                with self.assertRaisesRegex(
                        SystemError,
                        r"flags for %s must be " % member_name):
                    _testlimitedcapi.make_heaptype_with_member(
                        basicsize=-100,
                        add_relative_flag=True,
                        member_name=member_name,
                        member_offset=0,
                        member_type=_testlimitedcapi.Py_T_PYSSIZET,
                        member_flags=0,
                        )

    def test_pyobject_getitemdata_error(self):
        """Test PyObject_GetItemData fails on unsupported types"""
        with self.assertRaises(TypeError):
            # None is not variable-length
            _testcapi.pyobject_getitemdata(None)
        with self.assertRaises(TypeError):
            # int is variable-length, but doesn't have the
            # Py_TPFLAGS_ITEMS_AT_END layout (and flag)
            _testcapi.pyobject_getitemdata(0)


class TestPendingCalls(unittest.TestCase):

    # See the comment in ceval.c (at the "handle_eval_breaker" label)
    # about when pending calls get run.  This is especially relevant
    # here for creating deterministic tests.

    def main_pendingcalls_submit(self, l, n):
        def callback():
            #this function can be interrupted by thread switching so let's
            #use an atomic operation
            l.append(None)

        for i in range(n):
            time.sleep(random.random()*0.02) #0.01 secs on average
            #try submitting callback until successful.
            #rely on regular interrupt to flush queue if we are
            #unsuccessful.
            while True:
                if _testcapi._pending_threadfunc(callback):
                    break

    def pendingcalls_submit(self, l, n, *, main=True, ensure=False):
        def callback():
            #this function can be interrupted by thread switching so let's
            #use an atomic operation
            l.append(None)

        if main:
            return _testcapi._pending_threadfunc(callback, n,
                                                 blocking=False,
                                                 ensure_added=ensure)
        else:
            return _testinternalcapi.pending_threadfunc(callback, n,
                                                        blocking=False,
                                                        ensure_added=ensure)

    def pendingcalls_wait(self, l, numadded, context = None):
        #now, stick around until l[0] has grown to 10
        count = 0
        while len(l) != numadded:
            #this busy loop is where we expect to be interrupted to
            #run our callbacks.  Note that some callbacks are only run on the
            #main thread
            if False and support.verbose:
                print("(%i)"%(len(l),),)
            for i in range(1000):
                a = i*i
            if context and not context.event.is_set():
                continue
            count += 1
            self.assertTrue(count < 10000,
                "timeout waiting for %i callbacks, got %i"%(numadded, len(l)))
        if False and support.verbose:
            print("(%i)"%(len(l),))

    @threading_helper.requires_working_threading()
    def test_main_pendingcalls_threaded(self):

        #do every callback on a separate thread
        n = 32 #total callbacks
        threads = []
        class foo(object):pass
        context = foo()
        context.l = []
        context.n = 2 #submits per thread
        context.nThreads = n // context.n
        context.nFinished = 0
        context.lock = threading.Lock()
        context.event = threading.Event()

        threads = [threading.Thread(target=self.main_pendingcalls_thread,
                                    args=(context,))
                   for i in range(context.nThreads)]
        with threading_helper.start_threads(threads):
            self.pendingcalls_wait(context.l, n, context)

    def main_pendingcalls_thread(self, context):
        try:
            self.main_pendingcalls_submit(context.l, context.n)
        finally:
            with context.lock:
                context.nFinished += 1
                nFinished = context.nFinished
                if False and support.verbose:
                    print("finished threads: ", nFinished)
            if nFinished == context.nThreads:
                context.event.set()

    def test_main_pendingcalls_non_threaded(self):
        #again, just using the main thread, likely they will all be dispatched at
        #once.  It is ok to ask for too many, because we loop until we find a slot.
        #the loop can be interrupted to dispatch.
        #there are only 32 dispatch slots, so we go for twice that!
        l = []
        n = 64
        self.main_pendingcalls_submit(l, n)
        self.pendingcalls_wait(l, n)

    def test_max_pending(self):
        with self.subTest('main-only'):
            maxpending = 32

            l = []
            added = self.pendingcalls_submit(l, 1, main=True)
            self.pendingcalls_wait(l, added)
            self.assertEqual(added, 1)

            l = []
            added = self.pendingcalls_submit(l, maxpending, main=True)
            self.pendingcalls_wait(l, added)
            self.assertEqual(added, maxpending)

            l = []
            added = self.pendingcalls_submit(l, maxpending+1, main=True)
            self.pendingcalls_wait(l, added)
            self.assertEqual(added, maxpending)

        with self.subTest('not main-only'):
            # Per-interpreter pending calls has a much higher limit
            # on how many may be pending at a time.
            maxpending = 300

            l = []
            added = self.pendingcalls_submit(l, 1, main=False)
            self.pendingcalls_wait(l, added)
            self.assertEqual(added, 1)

            l = []
            added = self.pendingcalls_submit(l, maxpending, main=False)
            self.pendingcalls_wait(l, added)
            self.assertEqual(added, maxpending)

            l = []
            added = self.pendingcalls_submit(l, maxpending+1, main=False)
            self.pendingcalls_wait(l, added)
            self.assertEqual(added, maxpending)

    class PendingTask(types.SimpleNamespace):

        _add_pending = _testinternalcapi.pending_threadfunc

        def __init__(self, req, taskid=None, notify_done=None):
            self.id = taskid
            self.req = req
            self.notify_done = notify_done

            self.creator_tid = threading.get_ident()
            self.requester_tid = None
            self.runner_tid = None
            self.result = None

        def run(self):
            assert self.result is None
            self.runner_tid = threading.get_ident()
            self._run()
            if self.notify_done is not None:
                self.notify_done()

        def _run(self):
            self.result = self.req

        def run_in_pending_call(self, worker_tids):
            assert self._add_pending is _testinternalcapi.pending_threadfunc
            self.requester_tid = threading.get_ident()
            def callback():
                assert self.result is None
                # It can be tricky to control which thread handles
                # the eval breaker, so we take a naive approach to
                # make sure.
                if threading.get_ident() not in worker_tids:
                    self._add_pending(callback, ensure_added=True)
                    return
                self.run()
            self._add_pending(callback, ensure_added=True)

        def create_thread(self, worker_tids):
            return threading.Thread(
                target=self.run_in_pending_call,
                args=(worker_tids,),
            )

        def wait_for_result(self):
            while self.result is None:
                time.sleep(0.01)

    @threading_helper.requires_working_threading()
    def test_subthreads_can_handle_pending_calls(self):
        payload = 'Spam spam spam spam. Lovely spam! Wonderful spam!'

        task = self.PendingTask(payload)
        def do_the_work():
            tid = threading.get_ident()
            t = task.create_thread({tid})
            with threading_helper.start_threads([t]):
                task.wait_for_result()
        t = threading.Thread(target=do_the_work)
        with threading_helper.start_threads([t]):
            pass

        self.assertEqual(task.result, payload)

    @threading_helper.requires_working_threading()
    def test_many_subthreads_can_handle_pending_calls(self):
        main_tid = threading.get_ident()
        self.assertEqual(threading.main_thread().ident, main_tid)

        # We can't use queue.Queue since it isn't reentrant relative
        # to pending calls.
        _queue = deque()
        _active = deque()
        _done_lock = threading.Lock()
        def queue_put(task):
            _queue.append(task)
            _active.append(True)
        def queue_get():
            try:
                task = _queue.popleft()
            except IndexError:
                raise queue.Empty
            return task
        def queue_task_done():
            _active.pop()
            if not _active:
                try:
                    _done_lock.release()
                except RuntimeError:
                    assert not _done_lock.locked()
        def queue_empty():
            return not _queue
        def queue_join():
            _done_lock.acquire()
            _done_lock.release()

        tasks = []
        for i in range(20):
            task = self.PendingTask(
                req=f'request {i}',
                taskid=i,
                notify_done=queue_task_done,
            )
            tasks.append(task)
            queue_put(task)
        # This will be released once all the tasks have finished.
        _done_lock.acquire()

        def add_tasks(worker_tids):
            while True:
                if done:
                    return
                try:
                    task = queue_get()
                except queue.Empty:
                    break
                task.run_in_pending_call(worker_tids)

        done = False
        def run_tasks():
            while not queue_empty():
                if done:
                    return
                time.sleep(0.01)
            # Give the worker a chance to handle any remaining pending calls.
            while not done:
                time.sleep(0.01)

        # Start the workers and wait for them to finish.
        worker_threads = [threading.Thread(target=run_tasks)
                          for _ in range(3)]
        with threading_helper.start_threads(worker_threads):
            try:
                # Add a pending call for each task.
                worker_tids = [t.ident for t in worker_threads]
                threads = [threading.Thread(target=add_tasks, args=(worker_tids,))
                           for _ in range(3)]
                with threading_helper.start_threads(threads):
                    try:
                        pass
                    except BaseException:
                        done = True
                        raise  # re-raise
                # Wait for the pending calls to finish.
                queue_join()
                # Notify the workers that they can stop.
                done = True
            except BaseException:
                done = True
                raise  # re-raise
        runner_tids = [t.runner_tid for t in tasks]

        self.assertNotIn(main_tid, runner_tids)
        for task in tasks:
            with self.subTest(f'task {task.id}'):
                self.assertNotEqual(task.requester_tid, main_tid)
                self.assertNotEqual(task.requester_tid, task.runner_tid)
                self.assertNotIn(task.requester_tid, runner_tids)

    @requires_subinterpreters
    @support.skip_if_sanitizer("gh-129824: race on assign_version_tag", thread=True)
    def test_isolated_subinterpreter(self):
        # We exercise the most important permutations.

        # This test relies on pending calls getting called
        # (eval breaker tripped) at each loop iteration
        # and at each call.

        maxtext = 250
        main_interpid = 0
        interpid = _interpreters.create()
        self.addCleanup(lambda: _interpreters.destroy(interpid))
        _interpreters.run_string(interpid, f"""if True:
            import json
            import os
            import threading
            import time
            import _testinternalcapi
            from test.support import threading_helper
            """)

        def create_pipe():
            r, w = os.pipe()
            self.addCleanup(lambda: os.close(r))
            self.addCleanup(lambda: os.close(w))
            return r, w

        with self.subTest('add in main, run in subinterpreter'):
            r_ready, w_ready = create_pipe()
            r_done, w_done= create_pipe()
            timeout = time.time() + 30  # seconds

            def do_work():
                _interpreters.run_string(interpid, f"""if True:
                    # Wait until this interp has handled the pending call.
                    waiting = False
                    done = False
                    def wait(os_read=os.read):
                        global done, waiting
                        waiting = True
                        os_read({r_done}, 1)
                        done = True
                    t = threading.Thread(target=wait)
                    with threading_helper.start_threads([t]):
                        while not waiting:
                            pass
                        os.write({w_ready}, b'\\0')
                        # Loop to trigger the eval breaker.
                        while not done:
                            time.sleep(0.01)
                            if time.time() > {timeout}:
                                raise Exception('timed out!')
                    """)
            t = threading.Thread(target=do_work)
            with threading_helper.start_threads([t]):
                os.read(r_ready, 1)
                # Add the pending call and wait for it to finish.
                actual = _testinternalcapi.pending_identify(interpid)
                # Signal the subinterpreter to stop.
                os.write(w_done, b'\0')

            self.assertEqual(actual, int(interpid))

        with self.subTest('add in main, run in subinterpreter sub-thread'):
            r_ready, w_ready = create_pipe()
            r_done, w_done= create_pipe()
            timeout = time.time() + 30  # seconds

            def do_work():
                _interpreters.run_string(interpid, f"""if True:
                    waiting = False
                    done = False
                    def subthread():
                        while not waiting:
                            pass
                        os.write({w_ready}, b'\\0')
                        # Loop to trigger the eval breaker.
                        while not done:
                            time.sleep(0.01)
                            if time.time() > {timeout}:
                                raise Exception('timed out!')
                    t = threading.Thread(target=subthread)
                    with threading_helper.start_threads([t]):
                        # Wait until this interp has handled the pending call.
                        waiting = True
                        os.read({r_done}, 1)
                        done = True
                    """)
            t = threading.Thread(target=do_work)
            with threading_helper.start_threads([t]):
                os.read(r_ready, 1)
                # Add the pending call and wait for it to finish.
                actual = _testinternalcapi.pending_identify(interpid)
                # Signal the subinterpreter to stop.
                os.write(w_done, b'\0')

            self.assertEqual(actual, int(interpid))

        with self.subTest('add in subinterpreter, run in main'):
            r_ready, w_ready = create_pipe()
            r_done, w_done= create_pipe()
            r_data, w_data= create_pipe()
            timeout = time.time() + 30  # seconds

            def add_job():
                os.read(r_ready, 1)
                _interpreters.run_string(interpid, f"""if True:
                    # Add the pending call and wait for it to finish.
                    actual = _testinternalcapi.pending_identify({main_interpid})
                    # Signal the subinterpreter to stop.
                    os.write({w_done}, b'\\0')
                    os.write({w_data}, actual.to_bytes(1, 'little'))
                    """)
            # Wait until this interp has handled the pending call.
            waiting = False
            done = False
            def wait(os_read=os.read):
                nonlocal done, waiting
                waiting = True
                os_read(r_done, 1)
                done = True
            t1 = threading.Thread(target=add_job)
            t2 = threading.Thread(target=wait)
            with threading_helper.start_threads([t1, t2]):
                while not waiting:
                    pass
                os.write(w_ready, b'\0')
                # Loop to trigger the eval breaker.
                while not done:
                    time.sleep(0.01)
                    if time.time() > timeout:
                        raise Exception('timed out!')
                text = os.read(r_data, 1)
            actual = int.from_bytes(text, 'little')

            self.assertEqual(actual, int(main_interpid))

        with self.subTest('add in subinterpreter, run in sub-thread'):
            r_ready, w_ready = create_pipe()
            r_done, w_done= create_pipe()
            r_data, w_data= create_pipe()
            timeout = time.time() + 30  # seconds

            def add_job():
                os.read(r_ready, 1)
                _interpreters.run_string(interpid, f"""if True:
                    # Add the pending call and wait for it to finish.
                    actual = _testinternalcapi.pending_identify({main_interpid})
                    # Signal the subinterpreter to stop.
                    os.write({w_done}, b'\\0')
                    os.write({w_data}, actual.to_bytes(1, 'little'))
                    """)
            # Wait until this interp has handled the pending call.
            waiting = False
            done = False
            def wait(os_read=os.read):
                nonlocal done, waiting
                waiting = True
                os_read(r_done, 1)
                done = True
            def subthread():
                while not waiting:
                    pass
                os.write(w_ready, b'\0')
                # Loop to trigger the eval breaker.
                while not done:
                    time.sleep(0.01)
                    if time.time() > timeout:
                        raise Exception('timed out!')
            t1 = threading.Thread(target=add_job)
            t2 = threading.Thread(target=wait)
            t3 = threading.Thread(target=subthread)
            with threading_helper.start_threads([t1, t2, t3]):
                pass
            text = os.read(r_data, 1)
            actual = int.from_bytes(text, 'little')

            self.assertEqual(actual, int(main_interpid))

        # XXX We can't use the rest until gh-105716 is fixed.
        return

        with self.subTest('add in subinterpreter, run in subinterpreter sub-thread'):
            r_ready, w_ready = create_pipe()
            r_done, w_done= create_pipe()
            r_data, w_data= create_pipe()
            timeout = time.time() + 30  # seconds

            def do_work():
                _interpreters.run_string(interpid, f"""if True:
                    waiting = False
                    done = False
                    def subthread():
                        while not waiting:
                            pass
                        os.write({w_ready}, b'\\0')
                        # Loop to trigger the eval breaker.
                        while not done:
                            time.sleep(0.01)
                            if time.time() > {timeout}:
                                raise Exception('timed out!')
                    t = threading.Thread(target=subthread)
                    with threading_helper.start_threads([t]):
                        # Wait until this interp has handled the pending call.
                        waiting = True
                        os.read({r_done}, 1)
                        done = True
                    """)
            t = threading.Thread(target=do_work)
            #with threading_helper.start_threads([t]):
            t.start()
            if True:
                os.read(r_ready, 1)
                _interpreters.run_string(interpid, f"""if True:
                    # Add the pending call and wait for it to finish.
                    actual = _testinternalcapi.pending_identify({interpid})
                    # Signal the subinterpreter to stop.
                    os.write({w_done}, b'\\0')
                    os.write({w_data}, actual.to_bytes(1, 'little'))
                    """)
            t.join()
            text = os.read(r_data, 1)
            actual = int.from_bytes(text, 'little')

            self.assertEqual(actual, int(interpid))

    @threading_helper.requires_working_threading()
    def test_pending_call_creates_thread(self):
        source = """
        import _testinternalcapi
        import threading
        import time


        def output():
            print(24)
            time.sleep(1)
            print(42)


        def callback():
            threading.Thread(target=output).start()


        def create_pending_call():
            time.sleep(1)
            _testinternalcapi.simple_pending_call(callback)


        threading.Thread(target=create_pending_call).start()
        """
        return_code, stdout, stderr = script_helper.assert_python_ok('-c', textwrap.dedent(source))
        self.assertEqual(return_code, 0)
        self.assertEqual(stdout, f"24{os.linesep}42{os.linesep}".encode("utf-8"))
        self.assertEqual(stderr, b"")


class SubinterpreterTest(unittest.TestCase):

    @unittest.skipUnless(hasattr(os, "pipe"), "requires os.pipe()")
    def test_subinterps(self):
        import builtins
        r, w = os.pipe()
        code = """if 1:
            import sys, builtins, pickle
            with open({:d}, "wb") as f:
                pickle.dump(id(sys.modules), f)
                pickle.dump(id(builtins), f)
            """.format(w)
        with open(r, "rb") as f:
            ret = support.run_in_subinterp(code)
            self.assertEqual(ret, 0)
            self.assertNotEqual(pickle.load(f), id(sys.modules))
            self.assertNotEqual(pickle.load(f), id(builtins))

    @unittest.skipUnless(hasattr(os, "pipe"), "requires os.pipe()")
    def test_subinterps_recent_language_features(self):
        r, w = os.pipe()
        code = """if 1:
            import pickle
            with open({:d}, "wb") as f:

                @(lambda x:x)  # Py 3.9
                def noop(x): return x

                a = (b := f'1{{2}}3') + noop('x')  # Py 3.8 (:=) / 3.6 (f'')

                async def foo(arg): return await arg  # Py 3.5

                pickle.dump(dict(a=a, b=b), f)
            """.format(w)

        with open(r, "rb") as f:
            ret = support.run_in_subinterp(code)
            self.assertEqual(ret, 0)
            self.assertEqual(pickle.load(f), {'a': '123x', 'b': '123'})

    # _testcapi cannot be imported in a subinterpreter on a Free Threaded build
    @support.requires_gil_enabled()
    def test_py_config_isoloated_per_interpreter(self):
        # A config change in one interpreter must not leak to out to others.
        #
        # This test could verify ANY config value, it just happens to have been
        # written around the time of int_max_str_digits. Refactoring is okay.
        code = """if 1:
        import sys, _testcapi

        # Any config value would do, this happens to be the one being
        # double checked at the time this test was written.
        _testcapi.config_set('int_max_str_digits', 55555)
        sub_value = _testcapi.config_get('int_max_str_digits')
        assert sub_value == 55555, sub_value
        """
        before_config = _testcapi.config_get('int_max_str_digits')
        assert before_config != 55555
        self.assertEqual(support.run_in_subinterp(code), 0,
                         'subinterp code failure, check stderr.')
        after_config = _testcapi.config_get('int_max_str_digits')
        self.assertIsNot(
                before_config, after_config,
                "Expected get_config() to return a new dict on each call")
        self.assertEqual(before_config, after_config,
                         "CAUTION: Tests executed after this may be "
                         "running under an altered config.")
        # try:...finally: calling set_config(before_config) not done
        # as that results in sys.argv, sys.path, and sys.warnoptions
        # "being modified by test_capi" per test.regrtest.  So if this
        # test fails, assume that the environment in this process may
        # be altered and suspect.

    @requires_subinterpreters
    @unittest.skipUnless(hasattr(os, "pipe"), "requires os.pipe()")
    def test_configured_settings(self):
        """
        The config with which an interpreter is created corresponds
        1-to-1 with the new interpreter's settings.  This test verifies
        that they match.
        """

        OBMALLOC = 1<<5
        EXTENSIONS = 1<<8
        THREADS = 1<<10
        DAEMON_THREADS = 1<<11
        FORK = 1<<15
        EXEC = 1<<16
        ALL_FLAGS = (OBMALLOC | FORK | EXEC | THREADS | DAEMON_THREADS
                     | EXTENSIONS);

        features = [
            'obmalloc',
            'fork',
            'exec',
            'threads',
            'daemon_threads',
            'extensions',
            'own_gil',
        ]
        kwlist = [f'allow_{n}' for n in features]
        kwlist[0] = 'use_main_obmalloc'
        kwlist[-2] = 'check_multi_interp_extensions'
        kwlist[-1] = 'own_gil'

        expected_to_work = {
            (True, True, True, True, True, True, True):
                (ALL_FLAGS, True),
            (True, False, False, False, False, False, False):
                (OBMALLOC, False),
            (False, False, False, True, False, True, False):
                (THREADS | EXTENSIONS, False),
        }

        expected_to_fail = {
            (False, False, False, False, False, False, False),
        }

        # gh-117649: The free-threaded build does not currently allow
        # setting check_multi_interp_extensions to False.
        if Py_GIL_DISABLED:
            for config in list(expected_to_work.keys()):
                kwargs = dict(zip(kwlist, config))
                if not kwargs['check_multi_interp_extensions']:
                    del expected_to_work[config]
                    expected_to_fail.add(config)

        # expected to work
        for config, expected in expected_to_work.items():
            kwargs = dict(zip(kwlist, config))
            exp_flags, exp_gil = expected
            expected = {
                'feature_flags': exp_flags,
                'own_gil': exp_gil,
            }
            with self.subTest(config):
                r, w = os.pipe()
                script = textwrap.dedent(f'''
                    import _testinternalcapi, json, os
                    settings = _testinternalcapi.get_interp_settings()
                    with os.fdopen({w}, "w") as stdin:
                        json.dump(settings, stdin)
                    ''')
                with os.fdopen(r) as stdout:
                    ret = support.run_in_subinterp_with_config(script, **kwargs)
                    self.assertEqual(ret, 0)
                    out = stdout.read()
                settings = json.loads(out)

                self.assertEqual(settings, expected)

        # expected to fail
        for config in expected_to_fail:
            kwargs = dict(zip(kwlist, config))
            with self.subTest(config):
                script = textwrap.dedent(f'''
                    import _testinternalcapi
                    _testinternalcapi.get_interp_settings()
                    raise NotImplementedError('unreachable')
                    ''')
                with self.assertRaises(_interpreters.InterpreterError):
                    support.run_in_subinterp_with_config(script, **kwargs)

    @unittest.skipIf(_testsinglephase is None, "test requires _testsinglephase module")
    @unittest.skipUnless(hasattr(os, "pipe"), "requires os.pipe()")
    # gh-117649: The free-threaded build does not currently allow overriding
    # the check_multi_interp_extensions setting.
    @expected_failure_if_gil_disabled()
    def test_overridden_setting_extensions_subinterp_check(self):
        """
        PyInterpreterConfig.check_multi_interp_extensions can be overridden
        with PyInterpreterState.override_multi_interp_extensions_check.
        This verifies that the override works but does not modify
        the underlying setting.
        """

        OBMALLOC = 1<<5
        EXTENSIONS = 1<<8
        THREADS = 1<<10
        DAEMON_THREADS = 1<<11
        FORK = 1<<15
        EXEC = 1<<16
        BASE_FLAGS = OBMALLOC | FORK | EXEC | THREADS | DAEMON_THREADS
        base_kwargs = {
            'use_main_obmalloc': True,
            'allow_fork': True,
            'allow_exec': True,
            'allow_threads': True,
            'allow_daemon_threads': True,
            'own_gil': False,
        }

        def check(enabled, override):
            kwargs = dict(
                base_kwargs,
                check_multi_interp_extensions=enabled,
            )
            flags = BASE_FLAGS | EXTENSIONS if enabled else BASE_FLAGS
            settings = {
                'feature_flags': flags,
                'own_gil': False,
            }

            expected = {
                'requested': override,
                'override__initial': 0,
                'override_after': override,
                'override_restored': 0,
                # The override should not affect the config or settings.
                'settings__initial': settings,
                'settings_after': settings,
                'settings_restored': settings,
                # These are the most likely values to be wrong.
                'allowed__initial': not enabled,
                'allowed_after': not ((override > 0) if override else enabled),
                'allowed_restored': not enabled,
            }

            r, w = os.pipe()
            if Py_GIL_DISABLED:
                # gh-117649: The test fails before `w` is closed
                self.addCleanup(os.close, w)
            script = textwrap.dedent(f'''
                from test.test_capi.check_config import run_singlephase_check
                run_singlephase_check({override}, {w})
                ''')
            with os.fdopen(r) as stdout:
                ret = support.run_in_subinterp_with_config(script, **kwargs)
                self.assertEqual(ret, 0)
                out = stdout.read()
            results = json.loads(out)

            self.assertEqual(results, expected)

        self.maxDiff = None

        # setting: check disabled
        with self.subTest('config: check disabled; override: disabled'):
            check(False, -1)
        with self.subTest('config: check disabled; override: use config'):
            check(False, 0)
        with self.subTest('config: check disabled; override: enabled'):
            check(False, 1)

        # setting: check enabled
        with self.subTest('config: check enabled; override: disabled'):
            check(True, -1)
        with self.subTest('config: check enabled; override: use config'):
            check(True, 0)
        with self.subTest('config: check enabled; override: enabled'):
            check(True, 1)

    def test_mutate_exception(self):
        """
        Exceptions saved in global module state get shared between
        individual module instances. This test checks whether or not
        a change in one interpreter's module gets reflected into the
        other ones.
        """
        import binascii

        support.run_in_subinterp("import binascii; binascii.Error.foobar = 'foobar'")

        self.assertNotHasAttr(binascii.Error, "foobar")

    @unittest.skipIf(_testmultiphase is None, "test requires _testmultiphase module")
    # gh-117649: The free-threaded build does not currently support sharing
    # extension module state between interpreters.
    @expected_failure_if_gil_disabled()
    def test_module_state_shared_in_global(self):
        """
        bpo-44050: Extension module state should be shared between interpreters
        when it doesn't support sub-interpreters.
        """
        r, w = os.pipe()
        self.addCleanup(os.close, r)
        self.addCleanup(os.close, w)

        # Apple extensions must be distributed as frameworks. This requires
        # a specialist loader.
        if support.is_apple_mobile:
            loader = "AppleFrameworkLoader"
        else:
            loader = "ExtensionFileLoader"

        script = textwrap.dedent(f"""
            import importlib.machinery
            import importlib.util
            import os

            fullname = '_test_module_state_shared'
            origin = importlib.util.find_spec('_testmultiphase').origin
            loader = importlib.machinery.{loader}(fullname, origin)
            spec = importlib.util.spec_from_loader(fullname, loader)
            module = importlib.util.module_from_spec(spec)
            attr_id = str(id(module.Error)).encode()

            os.write({w}, attr_id)
            """)
        exec(script)
        main_attr_id = os.read(r, 100)

        ret = support.run_in_subinterp(script)
        self.assertEqual(ret, 0)
        subinterp_attr_id = os.read(r, 100)
        self.assertEqual(main_attr_id, subinterp_attr_id)

    @threading_helper.requires_working_threading()
    @unittest.skipUnless(hasattr(os, "pipe"), "requires os.pipe()")
    @requires_subinterpreters
    def test_pending_call_creates_thread_subinterpreter(self):
        interpreters = import_helper.import_module("concurrent.interpreters")
        r, w = os.pipe()
        source = f"""if True:
        import _testinternalcapi
        import threading
        import time
        import os


        def output():
            time.sleep(1)
            os.write({w}, b"x")
            os.close({w})


        def callback():
            threading.Thread(target=output).start()


        def create_pending_call():
            time.sleep(1)
            _testinternalcapi.simple_pending_call(callback)


        threading.Thread(target=create_pending_call).start()
        """
        interp = interpreters.create()
        interp.exec(source)
        interp.close()
        data = os.read(r, 1)
        self.assertEqual(data, b"x")
        os.close(r)


@requires_subinterpreters
class InterpreterConfigTests(unittest.TestCase):

    supported = {
        'isolated': types.SimpleNamespace(
            use_main_obmalloc=False,
            allow_fork=False,
            allow_exec=False,
            allow_threads=True,
            allow_daemon_threads=False,
            check_multi_interp_extensions=True,
            gil='own',
        ),
        'legacy': types.SimpleNamespace(
            use_main_obmalloc=True,
            allow_fork=True,
            allow_exec=True,
            allow_threads=True,
            allow_daemon_threads=True,
            check_multi_interp_extensions=bool(Py_GIL_DISABLED),
            gil='shared',
        ),
        'empty': types.SimpleNamespace(
            use_main_obmalloc=False,
            allow_fork=False,
            allow_exec=False,
            allow_threads=False,
            allow_daemon_threads=False,
            check_multi_interp_extensions=False,
            gil='default',
        ),
    }
    gil_supported = ['default', 'shared', 'own']

    def iter_all_configs(self):
        for use_main_obmalloc in (True, False):
            for allow_fork in (True, False):
                for allow_exec in (True, False):
                    for allow_threads in (True, False):
                        for allow_daemon in (True, False):
                            for checkext in (True, False):
                                for gil in ('shared', 'own', 'default'):
                                    yield types.SimpleNamespace(
                                        use_main_obmalloc=use_main_obmalloc,
                                        allow_fork=allow_fork,
                                        allow_exec=allow_exec,
                                        allow_threads=allow_threads,
                                        allow_daemon_threads=allow_daemon,
                                        check_multi_interp_extensions=checkext,
                                        gil=gil,
                                    )

    def assert_ns_equal(self, ns1, ns2, msg=None):
        # This is mostly copied from TestCase.assertDictEqual.
        self.assertEqual(type(ns1), type(ns2))
        if ns1 == ns2:
            return

        import difflib
        import pprint
        from unittest.util import _common_shorten_repr
        standardMsg = '%s != %s' % _common_shorten_repr(ns1, ns2)
        diff = ('\n' + '\n'.join(difflib.ndiff(
                       pprint.pformat(vars(ns1)).splitlines(),
                       pprint.pformat(vars(ns2)).splitlines())))
        diff = f'namespace({diff})'
        standardMsg = self._truncateMessage(standardMsg, diff)
        self.fail(self._formatMessage(msg, standardMsg))

    def test_predefined_config(self):
        def check(name, expected):
            expected = self.supported[expected]
            args = (name,) if name else ()

            config1 = _interpreters.new_config(*args)
            self.assert_ns_equal(config1, expected)
            self.assertIsNot(config1, expected)

            config2 = _interpreters.new_config(*args)
            self.assert_ns_equal(config2, expected)
            self.assertIsNot(config2, expected)
            self.assertIsNot(config2, config1)

        with self.subTest('default'):
            check(None, 'isolated')

        for name in self.supported:
            with self.subTest(name):
                check(name, name)

    def test_update_from_dict(self):
        for name, vanilla in self.supported.items():
            with self.subTest(f'noop ({name})'):
                expected = vanilla
                overrides = vars(vanilla)
                config = _interpreters.new_config(name, **overrides)
                self.assert_ns_equal(config, expected)

            with self.subTest(f'change all ({name})'):
                overrides = {k: not v for k, v in vars(vanilla).items()}
                for gil in self.gil_supported:
                    if vanilla.gil == gil:
                        continue
                    overrides['gil'] = gil
                    expected = types.SimpleNamespace(**overrides)
                    config = _interpreters.new_config(
                                                            name, **overrides)
                    self.assert_ns_equal(config, expected)

            # Override individual fields.
            for field, old in vars(vanilla).items():
                if field == 'gil':
                    values = [v for v in self.gil_supported if v != old]
                else:
                    values = [not old]
                for val in values:
                    with self.subTest(f'{name}.{field} ({old!r} -> {val!r})'):
                        overrides = {field: val}
                        expected = types.SimpleNamespace(
                            **dict(vars(vanilla), **overrides),
                        )
                        config = _interpreters.new_config(
                                                            name, **overrides)
                        self.assert_ns_equal(config, expected)

        with self.subTest('unsupported field'):
            for name in self.supported:
                with self.assertRaises(ValueError):
                    _interpreters.new_config(name, spam=True)

        # Bad values for bool fields.
        for field, value in vars(self.supported['empty']).items():
            if field == 'gil':
                continue
            assert isinstance(value, bool)
            for value in [1, '', 'spam', 1.0, None, object()]:
                with self.subTest(f'unsupported value ({field}={value!r})'):
                    with self.assertRaises(TypeError):
                        _interpreters.new_config(**{field: value})

        # Bad values for .gil.
        for value in [True, 1, 1.0, None, object()]:
            with self.subTest(f'unsupported value(gil={value!r})'):
                with self.assertRaises(TypeError):
                    _interpreters.new_config(gil=value)
        for value in ['', 'spam']:
            with self.subTest(f'unsupported value (gil={value!r})'):
                with self.assertRaises(ValueError):
                    _interpreters.new_config(gil=value)

    def test_interp_init(self):
        questionable = [
            # strange
            dict(
                allow_fork=True,
                allow_exec=False,
            ),
            dict(
                gil='shared',
                use_main_obmalloc=False,
            ),
            # risky
            dict(
                allow_fork=True,
                allow_threads=True,
            ),
            # ought to be invalid?
            dict(
                allow_threads=False,
                allow_daemon_threads=True,
            ),
            dict(
                gil='own',
                use_main_obmalloc=True,
            ),
        ]
        invalid = [
            dict(
                use_main_obmalloc=False,
                check_multi_interp_extensions=False
            ),
        ]
        if Py_GIL_DISABLED:
            invalid.append(dict(check_multi_interp_extensions=False))
        def match(config, override_cases):
            ns = vars(config)
            for overrides in override_cases:
                if dict(ns, **overrides) == ns:
                    return True
            return False

        def check(config):
            script = 'pass'
            rc = _testinternalcapi.run_in_subinterp_with_config(script, config)
            self.assertEqual(rc, 0)

        for config in self.iter_all_configs():
            if config.gil == 'default':
                continue
            if match(config, invalid):
                with self.subTest(f'invalid: {config}'):
                    with self.assertRaises(_interpreters.InterpreterError):
                        check(config)
            elif match(config, questionable):
                with self.subTest(f'questionable: {config}'):
                    check(config)
            else:
                with self.subTest(f'valid: {config}'):
                    check(config)

    def test_get_config(self):
        @contextlib.contextmanager
        def new_interp(config):
            interpid = _interpreters.create(config, reqrefs=False)
            try:
                yield interpid
            finally:
                try:
                    _interpreters.destroy(interpid)
                except _interpreters.InterpreterNotFoundError:
                    pass

        with self.subTest('main'):
            expected = _interpreters.new_config('legacy')
            expected.gil = 'own'
            if Py_GIL_DISABLED:
                expected.check_multi_interp_extensions = False
            interpid, *_ = _interpreters.get_main()
            config = _interpreters.get_config(interpid)
            self.assert_ns_equal(config, expected)

        with self.subTest('isolated'):
            expected = _interpreters.new_config('isolated')
            with new_interp('isolated') as interpid:
                config = _interpreters.get_config(interpid)
            self.assert_ns_equal(config, expected)

        with self.subTest('legacy'):
            expected = _interpreters.new_config('legacy')
            with new_interp('legacy') as interpid:
                config = _interpreters.get_config(interpid)
            self.assert_ns_equal(config, expected)

        with self.subTest('custom'):
            orig = _interpreters.new_config(
                'empty',
                use_main_obmalloc=True,
                gil='shared',
                check_multi_interp_extensions=bool(Py_GIL_DISABLED),
            )
            with new_interp(orig) as interpid:
                config = _interpreters.get_config(interpid)
            self.assert_ns_equal(config, orig)


@requires_subinterpreters
class InterpreterIDTests(unittest.TestCase):

    def add_interp_cleanup(self, interpid):
        def ensure_destroyed():
            try:
                _interpreters.destroy(interpid)
            except _interpreters.InterpreterNotFoundError:
                pass
        self.addCleanup(ensure_destroyed)

    def new_interpreter(self):
        id = _interpreters.create()
        self.add_interp_cleanup(id)
        return id

    def test_conversion_int(self):
        convert = _testinternalcapi.normalize_interp_id
        interpid = convert(10)
        self.assertEqual(interpid, 10)

    def test_conversion_coerced(self):
        convert = _testinternalcapi.normalize_interp_id
        class MyInt(str):
            def __index__(self):
                return 10
        interpid = convert(MyInt())
        self.assertEqual(interpid, 10)

    def test_conversion_from_interpreter(self):
        convert = _testinternalcapi.normalize_interp_id
        interpid = self.new_interpreter()
        converted = convert(interpid)
        self.assertEqual(converted, interpid)

    def test_conversion_bad(self):
        convert = _testinternalcapi.normalize_interp_id

        for badid in [
            object(),
            10.0,
            '10',
            b'10',
        ]:
            with self.subTest(f'bad: {badid!r}'):
                with self.assertRaises(TypeError):
                    convert(badid)

        badid = -1
        with self.subTest(f'bad: {badid!r}'):
            with self.assertRaises(ValueError):
                convert(badid)

        badid = 2**64
        with self.subTest(f'bad: {badid!r}'):
            with self.assertRaises(OverflowError):
                convert(badid)

    def test_lookup_exists(self):
        interpid = self.new_interpreter()
        self.assertTrue(
            _testinternalcapi.interpreter_exists(interpid))

    def test_lookup_does_not_exist(self):
        interpid = _testinternalcapi.unused_interpreter_id()
        self.assertFalse(
            _testinternalcapi.interpreter_exists(interpid))

    def test_lookup_destroyed(self):
        interpid = _interpreters.create()
        _interpreters.destroy(interpid)
        self.assertFalse(
            _testinternalcapi.interpreter_exists(interpid))

    def get_refcount_helpers(self):
        return (
            _testinternalcapi.get_interpreter_refcount,
            (lambda id: _interpreters.incref(id, implieslink=False)),
            _interpreters.decref,
        )

    def test_linked_lifecycle_does_not_exist(self):
        exists = _testinternalcapi.interpreter_exists
        is_linked = _testinternalcapi.interpreter_refcount_linked
        link = _testinternalcapi.link_interpreter_refcount
        unlink = _testinternalcapi.unlink_interpreter_refcount
        get_refcount, incref, decref = self.get_refcount_helpers()

        with self.subTest('never existed'):
            interpid = _testinternalcapi.unused_interpreter_id()
            self.assertFalse(
                exists(interpid))
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                is_linked(interpid)
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                link(interpid)
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                unlink(interpid)
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                get_refcount(interpid)
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                incref(interpid)
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                decref(interpid)

        with self.subTest('destroyed'):
            interpid = _interpreters.create()
            _interpreters.destroy(interpid)
            self.assertFalse(
                exists(interpid))
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                is_linked(interpid)
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                link(interpid)
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                unlink(interpid)
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                get_refcount(interpid)
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                incref(interpid)
            with self.assertRaises(_interpreters.InterpreterNotFoundError):
                decref(interpid)

    def test_linked_lifecycle_initial(self):
        is_linked = _testinternalcapi.interpreter_refcount_linked
        get_refcount, _, _ = self.get_refcount_helpers()

        # A new interpreter will start out not linked, with a refcount of 0.
        interpid = self.new_interpreter()
        linked = is_linked(interpid)
        refcount = get_refcount(interpid)

        self.assertFalse(linked)
        self.assertEqual(refcount, 0)

    def test_linked_lifecycle_never_linked(self):
        exists = _testinternalcapi.interpreter_exists
        is_linked = _testinternalcapi.interpreter_refcount_linked
        get_refcount, incref, decref = self.get_refcount_helpers()

        interpid = self.new_interpreter()

        # Incref will not automatically link it.
        incref(interpid)
        self.assertFalse(
            is_linked(interpid))
        self.assertEqual(
            1, get_refcount(interpid))

        # It isn't linked so it isn't destroyed.
        decref(interpid)
        self.assertTrue(
            exists(interpid))
        self.assertFalse(
            is_linked(interpid))
        self.assertEqual(
            0, get_refcount(interpid))

    def test_linked_lifecycle_link_unlink(self):
        exists = _testinternalcapi.interpreter_exists
        is_linked = _testinternalcapi.interpreter_refcount_linked
        link = _testinternalcapi.link_interpreter_refcount
        unlink = _testinternalcapi.unlink_interpreter_refcount

        interpid = self.new_interpreter()

        # Linking at refcount 0 does not destroy the interpreter.
        link(interpid)
        self.assertTrue(
            exists(interpid))
        self.assertTrue(
            is_linked(interpid))

        # Unlinking at refcount 0 does not destroy the interpreter.
        unlink(interpid)
        self.assertTrue(
            exists(interpid))
        self.assertFalse(
            is_linked(interpid))

    def test_linked_lifecycle_link_incref_decref(self):
        exists = _testinternalcapi.interpreter_exists
        is_linked = _testinternalcapi.interpreter_refcount_linked
        link = _testinternalcapi.link_interpreter_refcount
        get_refcount, incref, decref = self.get_refcount_helpers()

        interpid = self.new_interpreter()

        # Linking it will not change the refcount.
        link(interpid)
        self.assertTrue(
            is_linked(interpid))
        self.assertEqual(
            0, get_refcount(interpid))

        # Decref with a refcount of 0 is not allowed.
        incref(interpid)
        self.assertEqual(
            1, get_refcount(interpid))

        # When linked, decref back to 0 destroys the interpreter.
        decref(interpid)
        self.assertFalse(
            exists(interpid))

    def test_linked_lifecycle_incref_link(self):
        is_linked = _testinternalcapi.interpreter_refcount_linked
        link = _testinternalcapi.link_interpreter_refcount
        get_refcount, incref, _ = self.get_refcount_helpers()

        interpid = self.new_interpreter()

        incref(interpid)
        self.assertEqual(
            1, get_refcount(interpid))

        # Linking it will not reset the refcount.
        link(interpid)
        self.assertTrue(
            is_linked(interpid))
        self.assertEqual(
            1, get_refcount(interpid))

    def test_linked_lifecycle_link_incref_unlink_decref(self):
        exists = _testinternalcapi.interpreter_exists
        is_linked = _testinternalcapi.interpreter_refcount_linked
        link = _testinternalcapi.link_interpreter_refcount
        unlink = _testinternalcapi.unlink_interpreter_refcount
        get_refcount, incref, decref = self.get_refcount_helpers()

        interpid = self.new_interpreter()

        link(interpid)
        self.assertTrue(
            is_linked(interpid))

        incref(interpid)
        self.assertEqual(
            1, get_refcount(interpid))

        # Unlinking it will not change the refcount.
        unlink(interpid)
        self.assertFalse(
            is_linked(interpid))
        self.assertEqual(
            1, get_refcount(interpid))

        # Unlinked: decref back to 0 does not destroys the interpreter.
        decref(interpid)
        self.assertTrue(
            exists(interpid))
        self.assertEqual(
            0, get_refcount(interpid))


class TestStaticTypes(unittest.TestCase):

    _has_run = False

    @classmethod
    def setUpClass(cls):
        # The tests here don't play nice with our approach to refleak
        # detection, so we bail out in that case.
        if cls._has_run:
            raise unittest.SkipTest('these tests do not support re-running')
        cls._has_run = True

    @contextlib.contextmanager
    def basic_static_type(self, *args):
        cls = _testcapi.get_basic_static_type(*args)
        yield cls

    def test_pytype_ready_always_sets_tp_type(self):
        # The point of this test is to prevent something like
        # https://github.com/python/cpython/issues/104614
        # from happening again.

        # First check when tp_base/tp_bases is *not* set before PyType_Ready().
        with self.basic_static_type() as cls:
            self.assertIs(cls.__base__, object);
            self.assertEqual(cls.__bases__, (object,));
            self.assertIs(type(cls), type(object));

        # Then check when we *do* set tp_base/tp_bases first.
        with self.basic_static_type(object) as cls:
            self.assertIs(cls.__base__, object);
            self.assertEqual(cls.__bases__, (object,));
            self.assertIs(type(cls), type(object));


class TestThreadState(unittest.TestCase):

    @threading_helper.reap_threads
    @threading_helper.requires_working_threading()
    def test_thread_state(self):
        # some extra thread-state tests driven via _testcapi
        def target():
            idents = []

            def callback():
                idents.append(threading.get_ident())

            _testcapi._test_thread_state(callback)
            a = b = callback
            time.sleep(1)
            # Check our main thread is in the list exactly 3 times.
            self.assertEqual(idents.count(threading.get_ident()), 3,
                             "Couldn't find main thread correctly in the list")

        target()
        t = threading.Thread(target=target)
        t.start()
        t.join()

    @threading_helper.reap_threads
    @threading_helper.requires_working_threading()
    def test_thread_gilstate_in_clear(self):
        # See https://github.com/python/cpython/issues/119585
        class C:
            def __del__(self):
                _testcapi.gilstate_ensure_release()

        # Thread-local variables are destroyed in `PyThreadState_Clear()`.
        local_var = threading.local()

        def callback():
            local_var.x = C()

        _testcapi._test_thread_state(callback)

    @threading_helper.reap_threads
    @threading_helper.requires_working_threading()
    def test_gilstate_ensure_no_deadlock(self):
        # See https://github.com/python/cpython/issues/96071
        code = textwrap.dedent("""
            import _testcapi

            def callback():
                print('callback called')

            _testcapi._test_thread_state(callback)
            """)
        ret = assert_python_ok('-X', 'tracemalloc', '-c', code)
        self.assertIn(b'callback called', ret.out)

    def test_gilstate_matches_current(self):
        _testcapi.test_current_tstate_matches()


def get_test_funcs(mod, exclude_prefix=None):
    funcs = {}
    for name in dir(mod):
        if not name.startswith('test_'):
            continue
        if exclude_prefix is not None and name.startswith(exclude_prefix):
            continue
        funcs[name] = getattr(mod, name)
    return funcs


class Test_testcapi(unittest.TestCase):
    locals().update(get_test_funcs(_testcapi))

    # Suppress warning from PyUnicode_FromUnicode().
    @warnings_helper.ignore_warnings(category=DeprecationWarning)
    def test_widechar(self):
        _testlimitedcapi.test_widechar()

    def test_version_api_data(self):
        self.assertEqual(_testcapi.Py_Version, sys.hexversion)


class Test_testlimitedcapi(unittest.TestCase):
    locals().update(get_test_funcs(_testlimitedcapi))


class Test_testinternalcapi(unittest.TestCase):
    locals().update(get_test_funcs(_testinternalcapi,
                                   exclude_prefix='test_lock_'))


@threading_helper.requires_working_threading()
class Test_PyLock(unittest.TestCase):
    locals().update((name, getattr(_testinternalcapi, name))
                    for name in dir(_testinternalcapi)
                    if name.startswith('test_lock_'))


@unittest.skipIf(_testmultiphase is None, "test requires _testmultiphase module")
class Test_ModuleStateAccess(unittest.TestCase):
    """Test access to module start (PEP 573)"""

    # The C part of the tests lives in _testmultiphase, in a module called
    # _testmultiphase_meth_state_access.
    # This module has multi-phase initialization, unlike _testcapi.

    def setUp(self):
        fullname = '_testmultiphase_meth_state_access'  # XXX
        origin = importlib.util.find_spec('_testmultiphase').origin
        # Apple extensions must be distributed as frameworks. This requires
        # a specialist loader.
        if support.is_apple_mobile:
            loader = importlib.machinery.AppleFrameworkLoader(fullname, origin)
        else:
            loader = importlib.machinery.ExtensionFileLoader(fullname, origin)
        spec = importlib.util.spec_from_loader(fullname, loader)
        module = importlib.util.module_from_spec(spec)
        loader.exec_module(module)
        self.module = module

    def test_subclass_get_module(self):
        """PyType_GetModule for defining_class"""
        class StateAccessType_Subclass(self.module.StateAccessType):
            pass

        instance = StateAccessType_Subclass()
        self.assertIs(instance.get_defining_module(), self.module)

    def test_subclass_get_module_with_super(self):
        class StateAccessType_Subclass(self.module.StateAccessType):
            def get_defining_module(self):
                return super().get_defining_module()

        instance = StateAccessType_Subclass()
        self.assertIs(instance.get_defining_module(), self.module)

    def test_state_access(self):
        """Checks methods defined with and without argument clinic

        This tests a no-arg method (get_count) and a method with
        both a positional and keyword argument.
        """

        a = self.module.StateAccessType()
        b = self.module.StateAccessType()

        methods = {
            'clinic': a.increment_count_clinic,
            'noclinic': a.increment_count_noclinic,
        }

        for name, increment_count in methods.items():
            with self.subTest(name):
                self.assertEqual(a.get_count(), b.get_count())
                self.assertEqual(a.get_count(), 0)

                increment_count()
                self.assertEqual(a.get_count(), b.get_count())
                self.assertEqual(a.get_count(), 1)

                increment_count(3)
                self.assertEqual(a.get_count(), b.get_count())
                self.assertEqual(a.get_count(), 4)

                increment_count(-2, twice=True)
                self.assertEqual(a.get_count(), b.get_count())
                self.assertEqual(a.get_count(), 0)

                with self.assertRaises(TypeError):
                    increment_count(thrice=3)

                with self.assertRaises(TypeError):
                    increment_count(1, 2, 3)

    def test_get_module_bad_def(self):
        # PyType_GetModuleByDef fails gracefully if it doesn't
        # find what it's looking for.
        # see bpo-46433
        instance = self.module.StateAccessType()
        with self.assertRaises(TypeError):
            instance.getmodulebydef_bad_def()

    def test_get_module_static_in_mro(self):
        # Here, the class PyType_GetModuleByDef is looking for
        # appears in the MRO after a static type (Exception).
        # see bpo-46433
        class Subclass(BaseException, self.module.StateAccessType):
            pass
        self.assertIs(Subclass().get_defining_module(), self.module)


class TestInternalFrameApi(unittest.TestCase):

    @staticmethod
    def func():
        return sys._getframe()

    def test_code(self):
        frame = self.func()
        code = _testinternalcapi.iframe_getcode(frame)
        self.assertIs(code, self.func.__code__)

    def test_lasti(self):
        frame = self.func()
        lasti = _testinternalcapi.iframe_getlasti(frame)
        self.assertGreater(lasti, 0)
        self.assertLess(lasti, len(self.func.__code__.co_code))

    def test_line(self):
        frame = self.func()
        line = _testinternalcapi.iframe_getline(frame)
        firstline = self.func.__code__.co_firstlineno
        self.assertEqual(line, firstline + 2)


SUFFICIENT_TO_DEOPT_AND_SPECIALIZE = 100

class Test_Pep523API(unittest.TestCase):

    def do_test(self, func, names):
        actual_calls = []
        start = SUFFICIENT_TO_DEOPT_AND_SPECIALIZE
        count = start + SUFFICIENT_TO_DEOPT_AND_SPECIALIZE
        try:
            for i in range(count):
                if i == start:
                    _testinternalcapi.set_eval_frame_record(actual_calls)
                func()
        finally:
            _testinternalcapi.set_eval_frame_default()
        expected_calls = names * SUFFICIENT_TO_DEOPT_AND_SPECIALIZE
        self.assertEqual(len(expected_calls), len(actual_calls))
        for expected, actual in zip(expected_calls, actual_calls, strict=True):
            self.assertEqual(expected, actual)

    def test_inlined_binary_subscr(self):
        class C:
            def __getitem__(self, other):
                return None
        def func():
            C()[42]
        names = ["func", "__getitem__"]
        self.do_test(func, names)

    def test_inlined_call(self):
        def inner(x=42):
            pass
        def func():
            inner()
            inner(42)
        names = ["func", "inner", "inner"]
        self.do_test(func, names)

    def test_inlined_call_function_ex(self):
        def inner(x):
            pass
        def func():
            inner(*[42])
        names = ["func", "inner"]
        self.do_test(func, names)

    def test_inlined_for_iter(self):
        def gen():
            yield 42
        def func():
            for _ in gen():
                pass
        names = ["func", "gen", "gen", "gen"]
        self.do_test(func, names)

    def test_inlined_load_attr(self):
        class C:
            @property
            def a(self):
                return 42
        class D:
            def __getattribute__(self, name):
                return 42
        def func():
            C().a
            D().a
        names = ["func", "a", "__getattribute__"]
        self.do_test(func, names)

    def test_inlined_send(self):
        def inner():
            yield 42
        def outer():
            yield from inner()
        def func():
            list(outer())
        names = ["func", "outer", "outer", "inner", "inner", "outer", "inner"]
        self.do_test(func, names)


class Test_Pep523AllowSpecialization(unittest.TestCase):
    """Tests for _PyInterpreterState_SetEvalFrameFunc with
    allow_specialization=1."""

    def test_is_specialization_enabled_default(self):
        # With no custom eval frame, specialization should be enabled
        self.assertTrue(_testinternalcapi.is_specialization_enabled())

    def test_is_specialization_enabled_with_eval_frame(self):
        # Setting eval frame with allow_specialization=0 disables specialization
        try:
            _testinternalcapi.set_eval_frame_record([])
            self.assertFalse(_testinternalcapi.is_specialization_enabled())
        finally:
            _testinternalcapi.set_eval_frame_default()

    def test_is_specialization_enabled_after_restore(self):
        # Restoring the default eval frame re-enables specialization
        try:
            _testinternalcapi.set_eval_frame_record([])
            self.assertFalse(_testinternalcapi.is_specialization_enabled())
        finally:
            _testinternalcapi.set_eval_frame_default()
        self.assertTrue(_testinternalcapi.is_specialization_enabled())

    def test_is_specialization_enabled_with_allow(self):
        # Setting eval frame with allow_specialization=1 keeps it enabled
        try:
            _testinternalcapi.set_eval_frame_interp([])
            self.assertTrue(_testinternalcapi.is_specialization_enabled())
        finally:
            _testinternalcapi.set_eval_frame_default()

    def test_allow_specialization_call(self):
        def func():
            pass

        def func_outer():
            func()

        actual_calls = []
        try:
            _testinternalcapi.set_eval_frame_interp(
                actual_calls)
            for i in range(SUFFICIENT_TO_DEOPT_AND_SPECIALIZE * 2):
                func_outer()
        finally:
            _testinternalcapi.set_eval_frame_default()

        # With specialization enabled, calls to inner() will dispatch
        # through the installed frame evaluator
        self.assertEqual(actual_calls.count("func"), 0)

        # But the normal interpreter loop still shouldn't be inlining things
        self.assertNotEqual(actual_calls.count("func_outer"), 0)

    def test_no_specialization_call(self):
        # Without allow_specialization, ALL calls go through the eval frame.
        # This is the existing PEP 523 behavior.
        def inner(x=42):
            pass
        def func():
            inner()

        # Pre-specialize
        for _ in range(SUFFICIENT_TO_DEOPT_AND_SPECIALIZE):
            func()

        actual_calls = []
        try:
            _testinternalcapi.set_eval_frame_record(actual_calls)
            for _ in range(SUFFICIENT_TO_DEOPT_AND_SPECIALIZE):
                func()
        finally:
            _testinternalcapi.set_eval_frame_default()

        # Without allow_specialization, every call including inner() goes
        # through the eval frame
        expected = ["func", "inner"] * SUFFICIENT_TO_DEOPT_AND_SPECIALIZE
        self.assertEqual(actual_calls, expected)


@unittest.skipUnless(support.Py_GIL_DISABLED, 'need Py_GIL_DISABLED')
class TestPyThreadId(unittest.TestCase):
    def test_py_thread_id(self):
        # gh-112535: Test _Py_ThreadId(): make sure that thread identifiers
        # in a few threads are unique
        py_thread_id = _testinternalcapi.py_thread_id
        short_sleep = 0.010

        class GetThreadId(threading.Thread):
            def __init__(self):
                super().__init__()
                self.get_lock = threading.Lock()
                self.get_lock.acquire()
                self.started_lock = threading.Event()
                self.py_tid = None

            def run(self):
                self.started_lock.set()
                self.get_lock.acquire()
                self.py_tid = py_thread_id()
                time.sleep(short_sleep)
                self.py_tid2 = py_thread_id()

        nthread = 5
        threads = [GetThreadId() for _ in range(nthread)]

        # first make run sure that all threads are running
        for thread in threads:
            thread.start()
        for thread in threads:
            thread.started_lock.wait()

        # call _Py_ThreadId() in the main thread
        py_thread_ids = [py_thread_id()]

        # now call _Py_ThreadId() in each thread
        for thread in threads:
            thread.get_lock.release()

        # call _Py_ThreadId() in each thread and wait until threads complete
        for thread in threads:
            thread.join()
            py_thread_ids.append(thread.py_tid)
            # _PyThread_Id() should not change for a given thread.
            # For example, it should remain the same after a short sleep.
            self.assertEqual(thread.py_tid2, thread.py_tid)

        # make sure that all _Py_ThreadId() are unique
        for tid in py_thread_ids:
            self.assertIsInstance(tid, int)
            self.assertGreater(tid, 0)
        self.assertEqual(len(set(py_thread_ids)), len(py_thread_ids),
                         py_thread_ids)

class TestVersions(unittest.TestCase):
    full_cases = (
        (3, 4, 1, 0xA, 2, 0x030401a2),
        (3, 10, 0, 0xF, 0, 0x030a00f0),
        (0x103, 0x10B, 0xFF00, -1, 0xF0, 0x030b00f0),  # test masking
    )
    xy_cases = (
        (3, 4, 0x03040000),
        (3, 10, 0x030a0000),
        (0x103, 0x10B, 0x030b0000),  # test masking
    )

    def test_pack_full_version(self):
        for *args, expected in self.full_cases:
            with self.subTest(hexversion=hex(expected)):
                result = _testlimitedcapi.pack_full_version(*args)
                self.assertEqual(result, expected)

    def test_pack_version(self):
        for *args, expected in self.xy_cases:
            with self.subTest(hexversion=hex(expected)):
                result = _testlimitedcapi.pack_version(*args)
                self.assertEqual(result, expected)

    def test_pack_full_version_ctypes(self):
        ctypes = import_helper.import_module('ctypes')
        ctypes_func = ctypes.pythonapi.Py_PACK_FULL_VERSION
        ctypes_func.restype = ctypes.c_uint32
        ctypes_func.argtypes = [ctypes.c_int] * 5
        for *args, expected in self.full_cases:
            with self.subTest(hexversion=hex(expected)):
                result = ctypes_func(*args)
                self.assertEqual(result, expected)

    def test_pack_version_ctypes(self):
        ctypes = import_helper.import_module('ctypes')
        ctypes_func = ctypes.pythonapi.Py_PACK_VERSION
        ctypes_func.restype = ctypes.c_uint32
        ctypes_func.argtypes = [ctypes.c_int] * 2
        for *args, expected in self.xy_cases:
            with self.subTest(hexversion=hex(expected)):
                result = ctypes_func(*args)
                self.assertEqual(result, expected)


class TestCEval(unittest.TestCase):
   def test_ceval_decref(self):
        code = textwrap.dedent("""
            import _testcapi
            _testcapi.toggle_reftrace_printer(True)
            l1 = []
            l2 = []
            del l1
            del l2
            _testcapi.toggle_reftrace_printer(False)
        """)
        _, out, _ = assert_python_ok("-c", code)
        lines = out.decode("utf-8").splitlines()
        self.assertEqual(lines.count("CREATE list"), 2)
        self.assertEqual(lines.count("DESTROY list"), 2)


if __name__ == "__main__":
    unittest.main()