In a subinterpreter, spawning a daemon thread now raises an
exception. Daemon threads were never supported in subinterpreters.
Previously, the subinterpreter finalization crashed with a Pyton
fatal error if a daemon thread was still running.

* Add _thread._is_main_interpreter()
* threading.Thread.start() now raises RuntimeError if the thread is a
  daemon thread and the method is called from a subinterpreter.
* The _thread module now uses Argument Clinic for the new function.
* Use textwrap.dedent() in test_threading.SubinterpThreadingTests

Document reference cycle and resurrected objects issues in
sys.unraisablehook() and threading.excepthook() documentation.

Fix test.support.catch_unraisable_exception(): __exit__() no longer
ignores unraisable exceptions.

Fix test_io test_writer_close_error_on_close(): use a second
catch_unraisable_exception() to catch the BufferedWriter unraisable
exception.

This PR adds missing details in the [`concurrent.futures`](https://docs.python.org/3/library/concurrent.futures.html) documentation:

* the mention that `Future.cancel` also returns `False` if the call finished running;
* the mention of the states for `Future` that did not complete: pending or running.

Join the thread to prevent leaking a running thread and leaking a
reference.

Cleanup also the test:

* asyncioWindowsProactorEventLoopPolicy became the default policy,
  there is no need to set it manually.
* Only start the thread once the loop is running.
* Use a shorter sleep in the thread (100 ms rather than 1 sec).
* Use close_loop(loop) rather than loop.close().
* Use longer variable names.

Fix a regression introduced by af8646c8 that was causing code of the form:

if True and False:
   do_something()

to be optimized incorrectly, eliminating the block.

This is the followup  for issue36084



https://bugs.python.org/issue37077

* Replace global var Py_VerboseFlag with interp->config.verbose.
* Add _PyErr_NoMemory(tstate) function.
* Add tstate parameter to _PyEval_SetCoroutineOriginTrackingDepth()
  and move the function to the internal API.
* Replace _PySys_InitMain(runtime, interp)
  with _PySys_InitMain(runtime, tstate).

The peephole optimizer was not optimizing correctly bytecode after negative deltas were introduced. This is due to the fact that some special values (255) were being searched for in both instruction pointer delta and line number deltas.

It would raise ValueError("Paths don't have the same drive") if the paths on different drivers, which is not documented.


os.path.commonpath raises ValueError when the *paths* are in different drivers, but it is not documented.
Update the document according @Windsooon 's suggestion.
It actually raise ValueError according line 355 of [test of path](https://github.com/python/cpython/blob/master/Lib/test/test_ntpath.py) 


https://bugs.python.org/issue6689

The __exit__() method of test.support.catch_unraisable_exception
context manager now ignores unraisable exception raised when clearing
self.unraisable attribute.

Allow pure Python implementation of pickle to work
even when the C _pickle module is unavailable.

Fix test_pickle when _pickle is missing: declare PyPicklerHookTests
outside "if has_c_implementation:" block.

Remove the _tstate_lock from _shutdown_locks, don't remove None.

PyAST_obj2mod_ex() is similar to PyAST_obj2mod() with an additional
'feature_version' parameter which is unused.

GH-14039:  allow (no more than) one wholly empty arena on the usable_arenas list.

This prevents thrashing in some easily-provoked simple cases that could end up creating and destroying an arena on each loop iteration in client code.   Intuitively, if the only arena on the list becomes empty, it makes scant sense to give it back to the system unless we know we'll never need another free pool again before another arena frees a pool.  If the latter obtains, then - yes - this will "waste" an arena.

Remove ";" to fix Sphinx formatting.

Add a new _PyCompilerFlags_INIT macro to initialize PyCompilerFlags
variables, rather than initializing cf_flags and cf_feature_version
explicitly in each variable.

* Update PyCompilerFlags structure documentation.
* Document the new cf_feature_version field in the Changes in the C
  API section of the What's New in Python 3.8 doc.

Fix a race condition at Python shutdown when waiting for threads.
Wait until the Python thread state of all non-daemon threads get
deleted (join all non-daemon threads), rather than just wait until
Python threads complete.

* Add threading._shutdown_locks: set of Thread._tstate_lock locks
  of non-daemon threads used by _shutdown() to wait until all Python
  thread states get deleted. See Thread._set_tstate_lock().
* Add also threading._shutdown_locks_lock to protect access to
  threading._shutdown_locks.
* Add test_finalization_shutdown() test.

Calling setlocale(LC_CTYPE, "") on a system where GetACP() returns CP_UTF8 results in empty strings in _tzname[].

This causes time.tzname to be an empty string.
I have reported the bug to the UCRT team and will follow up, but it will take some time get a fix into production.

In the meantime one possible workaround is to temporarily change the locale by calling setlocale(LC_CTYPE, "C") before calling _tzset and restore the current locale after if the GetACP() == CP_UTF8 or CP_UTF7

@zooba 


https://bugs.python.org/issue36779

regrtest now uses sys.unraisablehook() to mark a test as "environment
altered" (ENV_CHANGED) if it emits an "unraisable exception".
Moreover, regrtest logs a warning in this case.

Use "python3 -m test --fail-env-changed" to catch unraisable
exceptions in tests.

Use catch_unraisable_exception() to ignore 'Exception ignored in:'
error when the internal BufferedWriter of the BufferedRWPair is
destroyed. The C implementation doesn't give access to the
internal BufferedWriter, so just ignore the warning instead.

because "getaddrinfo()" behaves different on AIX





https://bugs.python.org/issue35545

* bpo-29505: Enable fuzz testing of the json module, enforce size limit on int(x) fuzz and json input size to avoid timeouts.

Contributed by by Ammar Askar for Google.

Python 3.6 changed the size of bytecode instruction, while the documentation for `EXTENDED_ARG` was not updated accordingly.

Fix MSVC warning:

    objects\codeobject.c(285): warning C4244: '=':
    conversion from 'Py_ssize_t' to 'unsigned char',
    possible loss of data

Mock the HTTPConnection.close() method in a few unit tests to avoid
logging "Exception ignored in: ..." messages.

Rename compile() feature_version parameter to _feature_version and
convert it to a keyword-only parameter.

Update also test_type_comments to pass feature_version as a tuple.

(A single int is still allowed, but undocumented.)



https://bugs.python.org/issue35766