Constant.kind is added in https://bugs.python.org/issue36280.
Current possible values for Constant.kind are "u" or None.

For r'bar' and b'bar', Constant.kind value is None, so there's no need
for special handling.


https://bugs.python.org/issue37053

* No type cache for types with specialized mro, invalidation is hard.

* FIX: Don't disable method cache custom types that do not implement mro().

* fixing implem.

* Avoid storing error flags, also decref.

* news entry

* Clear as soon as we're getting an error.

* FIX: Reference leak.

The implementation is straightforward and essentially is just copied from `typing_extensions`.

The implementation is straightforward and essentially is just copied from `typing_extensions`.

The implementation is straightforward, it just mimics `ClassVar` (since the latter is also a name/access qualifier, not really a type). Also it is essentially copied from `typing_extensions`.

Update PyObject_CallMethodObjArgs and _PyObject_CallMethodIdObjArgs
to use _PyObject_GetMethod to avoid creating a bound method object
in many cases.

On a microbenchmark of PyObject_CallMethodObjArgs calling a method on
an interpreted Python class, this optimization resulted in a 1.7x
speedup.

* bpo-34632: Also include the test data directory.

* bpo-34632: remove the framework resources artifacts, accidentally added in 1bbf7b66 (ccbccce)

* 📝 Add a GitHub-specific security page

It will show up @
https://github.com/python/cpython/security/policy
allowing to navigate users who get there from "Security" tab in the
GitHub repo to the full article explaining the security vulnerability
reporting practices.
Co-Authored-By: Hugo <hugovk@users.noreply.github.com>

Add importlib.metadata module as forward port of the standalone importlib_metadata.

catch_unraisable_exception() now also removes its 'unraisable'
attribute at the context manager exit.

Add also missing AC_MSG_RESULT for AC_MSG_CHECKING(MACHDEP).

It is also possible to link against a library or executable with a
statically linked libpython, but not both with the same DLL.  In fact
building a statically linked python is currently broken on Cygwin
for other (related) reasons.

The same problem applies to other POSIX-like layers over Windows
(MinGW, MSYS) but Python's build system does not seem to attempt
to support those platforms at the moment.

* Fix a possible reference leak in _PyErr_Print() if exception
  is NULL.
* PyErr_BadInternalCall(): replace PyErr_Format() with _PyErr_SetString().
* Add pycore_pyerrors.h header file.
* New functions:

  * _PyErr_Clear()
  * _PyErr_Fetch()
  * _PyErr_Print()
  * _PyErr_Restore()
  * _PyErr_SetObject()
  * _PyErr_SetString()

* Add 'tstate' parameter to _PyEval_AddPendingCall().

PyOS_AfterFork_Child(): _PyInterpreterState_DeleteExceptMain() must
be called after _PyRuntimeState_ReInitThreads().

_PyRuntimeState_ReInitThreads() resets interpreters mutex after fork,
mutex used by _PyInterpreterState_DeleteExceptMain().

Add 'PyThreadState *tstate' parameter to errors.c functions to avoid
relying on global variables (indirectly on _PyRuntime).

Remove main_thread and main_interp variables from signalmodule.c:
reuse _PyRuntime which already track the main thread and the main
interpreter.

* Remove #include <sys/types.h> which became useless: getpid() call
  has been removed.
* Add runtime argument to is_main()
* is_main() now gets the interpreter from runtime.

Previously, it was hard to tell whether a function should be awaited. It was also incorrect (per PEP 484) to put this in the type hint for coroutine functions. Added this info to the output of builtins.help and pydoc.


https://bugs.python.org/issue36045

bpo-34626: Document creating heap types from the C-API

Add missing descriptions of PEP384's PyType_Spec and PyType_Slot,
along with some introductory prose.

``_thread.interrupt_main()`` now avoids setting the Python error status if the ``SIGINT`` signal is ignored or not handled by Python.

Adds sys.audit, sys.addaudithook, io.open_code, and associated C APIs.