* bpo-31249: Fix ref cycle in ThreadPoolExecutor

concurrent.futures: WorkItem.run() used by ThreadPoolExecutor now
breaks a reference cycle between an exception object and the WorkItem
object. ThreadPoolExecutor.shutdown() now also clears its threads
set.

* shutdown() now only clears threads if wait is true.

* Revert changes on shutdown()

PEP 523 introduced _PyEval_EvalFrameDefault which inlines PyEval_EvalFrameEx on
non-debug shared builds.  This breaks the ability to use py-bt, py-up, and
a few other Python-specific gdb integrations.

This patch fixes the problem by only looking for _PyEval_EvalFrameDefault
frames.

test_gdb passes on both a debug and a non-debug build.

Original patch by Bruno "Polaco" Penteado.

Title says all.

* Enforce that configparser defaults are strings
* Update test_configparser.py

bpo-30121: Test the Popen failure when Popen was created with pipes.
Create also NONEXISTING_CMD variable in test_subprocess.py.

* bpo-31238: pydoc ServerThread.stop() now joins itself

ServerThread.stop() now joins itself to wait until
DocServer.serve_until_quit() completes and then explicitly sets
its docserver attribute to None to break a reference cycle.

* Add NEWS.d entry

Add a BaseTestCase class to test_concurrent_futures to check for
dangling threads and processes on all tests, not only tests using
ExecutorMixin.

xmlrpc.server now explicitly breaks reference cycles when using
sys.exc_info() in code handling exceptions.

* bpo-9566: Silence warnings from pyatomic.h macros

Apparently MSVC is too stupid to understand that the alternate branch is
not taken and emits a warning for it.

Warnings added in https://github.com/python/cpython/pull/2383

* bpo-9566: A better fix for the pyatomic.h warning

* bpo-9566: Remove a slash

Part 3 of 3.  Remove old highlight functions and load_config as this functionality is now contained within classes. Patch by Cheryl Sabella.

Patch 2 of 3, to avoid horrendous diff.  Create highlights page from new HighPage class instead of old ConfigDialog methods and change tests to match.

test_bare_raise_in_brand_new_thread() now explicitly breaks a
reference cycle to not leak a dangling thread.

This is the first half of a patch similar to the one for for bpo-31205.  It is being split into 2 PRs to avoid what happened with PR-3096 -- an incomprehensible diff that could not be cleanly backported to 3.6.  This half copies several methods of ConfigDialog and turns them into a new class.

* bpo-30830: test_logging uses threading_setup/cleanup

Replace @support.reap_threads on some methods with
support.threading_setup() in setUp() and support.threading_cleanup()
in tearDown() in BaseTest.

* bpo-30830: test_logging disables threaded socketserver tests

Disable tests because of socketserver.ThreadingMixIn leaks threads,
whereas leaking threads now makes a test to fail on buildbots.

Disable tests until socketserver is fixed: bpo-31233.

* Skip also setup_via_listener()

* bpo-30947: Update libexpat from 2.2.1 to 2.2.3

* Add NEWS entry

* Add new loadlibrary.c

* expat_external.h: restore include "pyexpatns.h"

* PCbuild: add expat/loadlibrary.c

* Define XML_POOR_ENTROPY to compile expat

fork_wait.py tests now joins threads, to not leak running threads in
the background.

bpo-31231, bpo-30871: Replace "./python -m test.pythoninfo" with
"make pythoninfo", since macOS uses ./python.exe.

* bpo-30121: Fix debug assert in subprocess on Windows

This is caused by closing HANDLEs using os.close which is for CRT file
descriptors and not for HANDLEs.

* bpo-30121: Suppress debug assertion in test_subprocess when ran directly

* bpo-30871: pythoninfo: more sys, os, time data

PythonInfo now converts types other than intger to string by default.

* fix typo

The `Show Source` was broken because of a change made in sphinx 1.5.1
In Sphinx 1.4.9, the sourcename was "index.txt".
In Sphinx 1.5.1+, it is now "index.rst.txt"

Coroutines and async generators use a distinct attribute name for their
code objects, so this updates the `dis` module to correctly disassemble
objects with those attributes.

Due to the increase in the test module length, it also fixes some latent
defects in the tests related to how the displayed source line numbers
are extracted.

https://bugs.python.org/issue31230 is a follow-up issue suggesting we
may want to solve this a different way, by instead giving all these object
types a common `__code__` attribute, avoiding the need for special
casing in the `dis` module.

* bpo-30871: pythoninfo: add expat and _decimal

* Remove _decimal.__version__

The string is hardcoded, not really interesting.

* Add Lib/test/pythoninfo.py: script collecting various informations
  about Python to help debugging test failures.
* regrtest: remove sys.hash_info and sys.flags from header.
* Travis CI, Appveyor: run pythoninfo before tests

Tools/scripts/patchcheck.py now ignores changes in directories which
are copies of external libraries:

* Modules/_ctypes/libffi_msvc/
* Modules/_ctypes/libffi_osx/
* Modules/_decimal/libmpdec/
* Modules/expat/
* Modules/zlib/

Drop also support for Mercurial, since CPython migrated to Git.

Exclude also libmpdec

patchcheck: exclude also libffi_osx and libffi_msvc

Change the option for `Français` into `French` to be consistent with the other language selections that are already in English.

bpo-18966: non-daemonic threads created by a multiprocessing.Process should be joined on exit (#3111)

* bpo-18966: non-daemonic threads created by a multiprocessing.Process should be joined on exit

* Add NEWS blurb