- Issues #21936, #21163: Fix sporadic failures of
  test_future_exception_never_retrieved()
- Handle.cancel() now clears references to callback and args
- In debug mode, repr(Handle) now contains the location where the Handle was
  created.

until protocol.connection_made() has been called. Document also why transport
constructors use a waiter.

- Document the new create_task() method
- "Hide" the Task class: point to the create_task() method for interoperability
- Rewrite the documentation of the Task class
- Document the "Pending task destroyed"
- Update output in debug mode of examples in the dev section
- Replace Task() with create_task() in examples

- Tulip issue 185: Add a create_task() method to event loops. The create_task()
  method can be overriden in custom event loop to implement their own task
  class. For example, greenio and Pulsar projects use their own task class. The
  create_task() method is now preferred over creating directly task using the
  Task class.
- tests: fix a warning
- fix typo in the name of a test function
- Update AbstractEventLoop: add new event loop methods; update also the unit test

the number of received bytes is negative.

get a bytes string

PySys_FormatStderr() call

size to be a Py_ssize_t, not an int. _ssl.c is now "Py_ssize_t clean".

Initial patch by Kathleen Weaver.

Backout the "Tulip issue 181: Faster create_connection()" changeset, it was a
mistake.

- Tulip issue #181: Faster create_connection(). Call directly
  waiter.set_result() in the constructor of _ProactorBasePipeTransport and
  _SelectorSocketTransport, instead of using of delaying the call with
  call_soon().
- Cleanup iscoroutine()

representations (on mips and m68k platforms).

Patch by July Tikhonov.

Issue #21714: Disallow the construction of invalid paths using Path.with_name().  Original patch by Antony Lee.

due to possible uninitialized _config_vars.  Original patch by Alex Gaynor.

loop at exit. Patch written by Vajrasky Kok (I modified also the "hello world"
example using a coroutine).

of a Future with call_soon(). Add an helper, a private method, to set the
result only if the future was not cancelled.

Patch by Stefan Tatschner.

Issue #21897: Fix a crash with the f_locals attribute with closure variables when frame.clear() has been called.

uses "%zd" for Py_ssize_t, not PY_FORMAT_SIZE_T

Reported by Draic Kin on docs@.

When ``None`` was passed as a ``REG_BINARY`` value to SetValueEx,
PyMem_DEL was called on an uninitialized buffer.  Patch by John Ehresman.

(Also an incidental typo fix in a comment in test_winreg)

* _UnixSubprocessTransport: fix file mode of stdin. Open stdin in write mode,
  not in read mode
* Examples: close the event loop at exit
* More reliable CoroWrapper.__del__. If the constructor is interrupted by
  KeyboardInterrupt or the coroutine objet is destroyed lately, some the
  _source_traceback attribute doesn't exist anymore.
* repr(Task): include also the future the task is waiting for

it ignored I/O errors if at least the first C call read() succeed.