Declare functions with EXTINLINE:

* mpd_del()
* mpd_uint_zero()
* mpd_qresize()
* mpd_qresize_zero()
* mpd_minalloc()

These functions are implemented with "inline" or "ALWAYS_INLINE", but
declared without inline which cause linker error on Visual Studio in
Debug mode when using /Ob1.

Prior to this revision, after the shutdown of a `BaseServer`,
the server accepted a last single request
if it was sent between the server socket polling
and the polling timeout.

This can be problematic for instance for a server restart
for which you do not want to interrupt the service,
by not closing the listening socket during the restart.
One request failed because of this behavior.

Note that only one request failed,
following requests were not accepted, as expected.

Visual Studio solution: Set InlineFunctionExpansion to
OnlyExplicitInline ("/Ob1" option) on all projects (in
pyproject.props) in Debug mode on Win32 and x64 platforms to expand
functions marked as inline.

This change should make Python compiled in Debug mode a little bit
faster on Windows. On Unix, GCC uses -Og optimization level for
./configure --with-pydebug.

* Convert PyObject_INIT() and PyObject_INIT_VAR() macros to static
  inline functions.
* Fix usage of these functions: cast to PyObject* or PyVarObject*.

bpo-34890: Make iscoroutinefunction, isgeneratorfunction and isasyncgenfunction work with functools.partial (GH-9903)

inspect.isfunction() processes both inspect.isfunction(func) and
inspect.isfunction(partial(func, arg)) correctly but some other functions in the
inspect module (iscoroutinefunction, isgeneratorfunction and isasyncgenfunction)
lack this functionality. This commits adds a new check in the mentioned functions
in the inspect module so they can work correctly with arbitrarily nested partial
functions. 

_Py_NegativeRefcount() now uses _PyObject_AssertFailed() to dump the
object to help debugging.

_PyTraceMalloc_NewReference() is now called by _Py_NewReference(), so
move its definition to object.h. Moreover, define it even if
Py_LIMITED_API is defined, since _Py_NewReference() is also exposed
even if Py_LIMITED_API is defined.

The MagicMock class supports many magic methods, but not __fspath__. To ease
testing with modules such as os.path, this function is now supported by default.

https://bugs.python.org/issue35038

Changes:

* Add _PyObject_AssertFailed() function.
* Add _PyObject_ASSERT() and _PyObject_ASSERT_WITH_MSG() macros.
* gc_decref(): replace assert() with _PyObject_ASSERT_WITH_MSG() to
  dump the faulty object if the assertion fails.

_PyObject_AssertFailed() calls:

* _PyMem_DumpTraceback(): try to log the traceback where the object
  memory has been allocated if tracemalloc is enabled.
* _PyObject_Dump(): log repr(obj).
* Py_FatalError(): log the current Python traceback.

_PyObject_AssertFailed() uses _PyObject_IsFreed() heuristic to check
if the object memory has been freed by a debug hook on Python memory
allocators.

Initial patch written by David Malcolm.
Co-Authored-By: David Malcolm <dmalcolm@redhat.com>

* Add Py_STATIC_INLINE() macro to declare a "static inline" function.
  If the compiler supports it, try to always inline the function even if no
  optimization level was specified.
* Modify pydtrace.h to use Py_STATIC_INLINE() when WITH_DTRACE is
  not defined.
* Add an unit test on Py_DECREF() to make sure that
  _Py_NegativeRefcount() reports the correct filename.

Regression in b0bf51b3.

* Modify object.h to ensure that pymem.h is included,
  to get _Py_tracemalloc_config variable.
* Move _PyTraceMalloc_XXX() functions to tracemalloc.h,
  they need PyObject type. Break circular dependency between pymem.h
  and object.h.

tracemalloc now tries to update the traceback when an object is
reused from a "free list" (optimization for faster object creation,
used by the builtin list type for example).

Changes:

* Add _PyTraceMalloc_NewReference() function which tries to update
  the Python traceback of a Python object.
* _Py_NewReference() now calls _PyTraceMalloc_NewReference().
* Add an unit test.

.o generated by clang in LTO mode actually are LLVM bitcode files, which
leads to a few errors during configure/build step:

- add lto flags to the BASECFLAGS instead of CFLAGS, as CFLAGS are used
  to build autoconf test case, and some are not compatible with clang LTO
  (they assume binary in the .o, not bitcode)
- force llvm-ar instead of ar, as ar is not aware of .o files generated
  by clang -flto

The spec gets stored on modules with the __spec__ attribute, not spec.

Actually time calculations were remove in commit 891a1f86.

Replace str.format with f-strings in the code examples of asyncio-task documentation.

They no longer cache the wchar_t* representation of string objects.

Based on the investigation by Xiang Zhang.

References could leak, NULL could be dereferenced, and the Expat parser could
be double freed when some errors raised.

_PyObject_Dump() now uses an heuristic to check if the object memory
has been freed: log "<freed object>" in that case.

The heuristic rely on the debug hooks on Python memory allocators
which fills the memory with DEADBYTE (0xDB) when memory is
deallocated. Use PYTHONMALLOC=debug to always enable these debug
hooks.

Fix the documentation of copy2, as it does not copy file ownership (user and
group), only mode, mtime, atime and flags.

The original text was confusing to developers as it suggested that this
command is the same as 'cp -p', but according to cp(1), '-p' copies file
ownership as well.

Clarify which metadata is copied by shutil.copystat in its docstring.

https://bugs.python.org/issue35028