Include/*.h should be the "portable Python API", whereas
Include/cpython/*.h should be the "CPython API": CPython
implementation details.

Changes:

* Create Include/cpython/ subdirectory
* "make install" now creates $prefix/include/cpython and copy
  Include/cpython/* to $prefix/include/cpython
* Create Include/cpython/objimpl.h: move objimpl.h code
  surrounded by "#ifndef Py_LIMITED_API" to cpython/objimpl.h.
* objimpl.h now includes cpython/objimpl.h
* Windows installer (MSI) now also install Include/ subdirectories:
  Include/cpython/ and Include/internal/.

Modify the following fnctl function to retry if interrupted by a signal
(EINTR): flock, lockf, fnctl.

make install now also installs the internal API: Include/internal/*.h
header files.

That section is a tip on how to kill process on Windows for Python prior to 2.7 and 3.2.
3.1 end of support was April 2012 and 2.6 was October 2013, so that hasn't been need for supported versions of Python for more than 5 years. Beside not being needed anymore for a long time, when I read it with the eyes of a Python profane, it makes Python looks bad, like a language from the parts with warts you need to circumvent.
Let's delete that :)

PyObject_INIT() and PyObject_INIT_VAR() now cast their first argument
to PyObject*, as done in Python 3.7.

Revert partially commit b4435e20.

bpo-34523, bpo-35290: C locale coercion now resets the Python
internal "force ASCII" mode. This change fix the filesystem encoding
on FreeBSD CURRENT, which has a new "C.UTF-8" locale, when
the UTF-8 mode is disabled.

Add _Py_ResetForceASCII(): _Py_SetLocaleFromEnv() now calls it.

Add pycore_fileutils.h and pycore_object.h to Makefile.pre.in and to
the pythoncore project of PCbuild/.

Partially revert commit 5f2df88b:
add "#undef Yield" to .c files after including Python-ast.h.

Fix the warning:

    winbase.h(102): warning C4005: 'Yield': macro redefinition

If _PyObject_Dump() detects that the object is freed, don't try to
dump it (exit immediately).

Enhance also _PyObject_IsFreed(): it now detects if the pointer
itself looks like freed memory.

os_read_impl() now also truncates the size to _PY_READ_MAX
on macOS, to avoid to allocate a larger buffer even if _Py_read() is
limited to _PY_READ_MAX bytes (ex: INT_MAX on macOS).

When iterating using asdl_seq_LEN(), use 'Py_ssize_t' type instead of
'int' for the iterator variable, to avoid downcast on 64-bit platforms.

_Py_asdl_int_seq_new() now also ensures that the index is greater than
or equal to 0.

CryptGenRandom() maximum size is PY_DWORD_MAX, not INT_MAX.
Use DWORD type for the 'chunk' variable
Co-Authored-By: Jeremy Kloth <jeremy.kloth@gmail.com>

Explicit cast a pointer difference (intptr_t) to int to fix
two warnings on 64-bit Windows:

    Modules\pyexpat.c(1181): warning C4244: 'initializing':
    conversion from '__int64' to 'int', possible  loss of data

    Modules\pyexpat.c(1192): warning C4244: 'initializing':
    conversion from '__int64' to 'int', possible  loss of data

Fixed the following compiler warning in multibytecodec.c:

    warning C4244: '=': conversion from 'Py_ssize_t'
    to 'unsigned char', possible loss of data

Cast Py_ssize_t to unsigned char: the maximum value is checked
on the previous line.

Don't pass void* to Python macros: use _PyObject_CAST().

Don't pass complex expressions but regular variables to Python
macros.

* _datetimemodule.c: split single large "if" into two "if"
  in date_new(), time_new() and datetime_new().
* _pickle.c, load_extension(): flatten complex "if" expression into
  more regular C code.
* _ssl.c: addbool() now uses a temporary bool_obj to only evaluate
  the value once.
* weakrefobject.c: replace "Py_INCREF(result = proxy);"
  with "result = proxy; Py_INCREF(result);"

Add _PyObject_CAST() and _PyVarObject_CAST() macros to cast argument
to PyObject* and PyVarObject* properly.

* Add _PyObject_ASSERT_FROM() and _PyObject_ASSERT_FAILED_MSG()
  macros.
* PyObject_GC_Track() now calls _PyObject_ASSERT_FAILED_MSG(),
  instead of Py_FatalError(), if the object is already tracked, to
  dump more information on error.
* _PyObject_GC_TRACK() no longer checks if the object is already
  tracked at runtime, use an assertion instead for best performances;
  PyObject_GC_Track() still checks at runtime.
* pycore_object.h now includes pycore_pystate.h.
* Convert _PyObject_GC_TRACK() and _PyObject_GC_UNTRACK() macros to
  inline functions.

Enhance _PyObject_AssertFailed()

* Exchange 'expr' and 'msg' parameters
* 'expr' and 'func' arguments can now be NULL

Move _PyObject_GC_TRACK() and _PyObject_GC_UNTRACK() from
Include/objimpl.h to Include/internal/pycore_object.h.

Call setitimer() before each test method, instead of once per test
case, to ensure that signals are sent in each test method.
Previously, only the first method of a testcase class got signals.

Changes:

* Replace setUpClass() with setUp() and replace tearDownClass() with
  tearDown().
* tearDown() now ensures that at least one signal has been sent.
* Replace support.run_unittest() with unittest.main() which has
  a nicer CLI and automatically discover test cases.

I'll watch for 404 on the old URL and will setup an HTTP redirection if needed.

In verbose mode, test_c_locale_coercion now dumps global variables at
startup.

Partially revert commit 1a6be91e,
move back PyGC API from the internal API to the C API:

* _PyGCHead_NEXT(g), _PyGCHead_SET_NEXT(g, p)
* _PyGCHead_PREV(g), _PyGCHead_SET_PREV(g, p)
* _PyGCHead_FINALIZED(g), _PyGCHead_SET_FINALIZED(g)
* _PyGC_FINALIZED(o), _PyGC_SET_FINALIZED(o)
* _PyGC_PREV_MASK_FINALIZED
* _PyGC_PREV_MASK_COLLECTING
* _PyGC_PREV_SHIFT
* _PyGC_PREV_MASK

_PyObject_GC_TRACK(o) and _PyObject_GC_UNTRACK(o) remain in the
internal API.

Fixes assertion failures in _datetimemodule.c
introduced in the previous fix (see bpo-31752).

Rather of trying to handle an int subclass as exact int,
let it to use overridden special methods, but check the
result of divmod().

locale.localeconv() now sets temporarily the LC_CTYPE locale to the
LC_MONETARY locale if the two locales are different and monetary
strings are non-ASCII. This temporary change affects other threads.

Changes:

* locale.localeconv() can now set LC_CTYPE to LC_MONETARY to decode
  monetary fields.
* Add LocaleInfo.grouping_buffer: copy localeconv() grouping string
  since it can be replaced anytime if a different thread calls
  localeconv().
* _Py_GetLocaleconvNumeric() now requires a "struct lconv *"
  structure, so locale.localeconv() now longer calls localeconv()
  twice. Moreover, the function now requires all arguments to be
  non-NULL.
* Rename STATIC_LOCALE_INFO_INIT to LocaleInfo_STATIC_INIT.
* Move _Py_GetLocaleconvNumeric() definition from fileutils.h
  to pycore_fileutils.h. pycore_fileutils.h now includes locale.h.
* The _locale module is now built with Py_BUILD_CORE defined.

coro->cr_origin wasn't initialized if compute_cr_origin() failed in
PyCoro_New(), which would cause a crash during the coroutine's
deallocation.



https://bugs.python.org/issue35269

Source of T_STRING: https://github.com/python/cpython/blob/e42b705188271da108de42b55d9344642170aa2b/Python/structmember.c#L51

Source of PyUnicode_FromString
https://github.com/python/cpython/blob/master/Include/unicodeobject.h#L702


https://bugs.python.org/issue25438