Implement PEP 553, built-in breakpoint() with support from sys.breakpointhook(), along with documentation and tests.  Closes bpo-31353

Fix test_exception_errpipe_bad_data() and
test_exception_errpipe_normal() of test_subprocess: mock os.waitpid()
to avoid calling the real os.waitpid(0, 0) which is an unexpected
side effect of the test.

sre_compile does bit test (e.g. `flags & SRE_FLAG_IGNORECASE`) in loop.
`IntFlag.__and__` and `IntFlag.__new__` made it slower.

So this commit convert it to normal int before passing flags to `sre_compile()`.

Followed format of other entries.

Implementations of these operations are virtually identical.

* Fix incorrect links.
* Remove redundant links.
* Add signatures and index entries for gettext related
functions in the locale module.

On Windows, Py_FatalError() now limits the size to 256 bytes of the
buffer used to call OutputDebugStringW(). Previously, the size
depended on the length of the error message.

While there is not real bug in this case, using re.IGNORECASE without re.ASCII
leads unexpected behavior.
Instead of adding re.ASCII, this commit removes re.IGNORECASE flag because
it's easier and simpler.

This commit removes dead copy of the pattern in email.util module too.
While the pattern is same, it is compiled separately because it had different flags.

when pass a string larger than 2 GiB.

Decrease memory requirements for Tcl's bigmem tests.

Update the compile tests for optimization levels to also check that
__debug__ blocks are included or excluded based on the optimization
level.
Patch by Diana Clarke.

to integer with binary base.

It shows show import time of each module.
It's useful for optimizing startup time.

Typical usage: python -X importtime -c 'import requests'

On macOS, a process can exit with -SIGKILL if it is killed "early"
with SIGTERM.

On macOS, use uuid_generate_time() instead of
uuid_generate_time_safe() of libuuid, since uuid_generate_time_safe()
is not available.

bpo-31516: current_thread() should not return a dummy thread at shutdown

* bpo-31158: Fix nondeterministic read in test_pty

* Reuse existing readline implementation from io.

Thx to @pitrou

* Updated comment

Ideally, this commit is fixuped into the previous commit. Since there is
already a comment on github, I won't rebase.

DER_cert_to_PEM_cert() used textwrap.fill() to format PEM.
But it's library to wrap lines on word boundary, while PEM is
base64 encoded string.

Additionally, importing textwrap is little slow.

Save a bit of code, plus non-tuples get converted anyway to tuples by tkinter for the tk call.

Speed up class creation by 10-20% by reducing the overhead in the
necessary special method lookups.

Passing a widget instead of an flist with a root widget opens the option of
creating a browser frame that is only part of a window. Passing a full file
name instead of pieces assumed to come from a .py file opens the possibility
of browsing python files that do not end in .py.

bpo-31592: Fix an assertion failure in Python parser in case of a bad unicodedata.normalize(). (#3767)

* Allow whitespace inside of ElementPath predicates.
* Add ElementPath predicate support for text comparison of the current node, like "[.='text']".

netrc file format doesn't support quotes and escapes.

See https://linux.die.net/man/5/netrc

Importlib was instrumented with two dtrace probes to profile import timing.
Signed-off-by: Christian Heimes <christian@python.org>

This was possible before.  GH-1560 introduced a regression after 3.6.2 got
released where only sequences were accepted now.  This commit addresses this
problem.