Makes the documentation of math and cmath module
more helpful for the beginners.

highlightlang is deprecated since April 2018 in Sphinx.
See https://github.com/sphinx-doc/sphinx/pull/4845

Use _PyCoreConfig.program_name instead.

This PR proposes a solution to [bpo-35545](https://bugs.python.org/issue35545) by adding an optional `flowinfo` and `scopeid` to `asyncio.base_events._ipaddr_info` to carry the full address information into `_ipaddr_info` and avoid discarding IPv6 specific information.

Changelog entry & regression tests to come.


https://bugs.python.org/issue35545

The final addition (cur += step) may overflow, so use size_t for "cur".
"cur" is always positive (even for negative steps), so it is safe to use
size_t here.
Co-Authored-By: Martin Panter <vadmium+py@gmail.com>

precmdline is only needed in _PyCoreConfig_Read(), not in
config_read_cmdline(). Reorganize _PyCoreConfig_Read().

Add tests for configure_c_stdio and pathconfig_warnings parameters.

* Add a private _Py_InitializeMain() function.
* Add again _PyCoreConfig._init_main.
* _Py_InitializeFromConfig() now uses _init_main to decide
  if _Py_InitializeMainInterpreter() should be called.
* _PyCoreConfig: rename _frozen to pathconfig_warnings, its value is
  now the opposite of Py_FrozenFlag.
* Add an unit test for _init_main=0 and _Py_InitializeMain().

* _PyCoreConfig_Read() doesn't parse nor update argv
  if parse_argv is 0.
* Move path configuration fields in _PyCoreConfig.
* Add an unit test for parse_argv=0.
* Remove unused "done": label in _Py_RunMain().

The second attempt. Now deprecate `@coroutine` only, keep `yield from fut` as is.


https://bugs.python.org/issue36921

We cannot use "unsigned int" for exitcode on Windows, since
Py_Main() and _Py_RunMain() always return an "int".

Changes:

* _PyPathConfig_ComputeSysPath0() now returns -1 if an exception is
  raised.
* pymain_run_python() no longer uses _PyInitError but display the
  exception and set exitcode to 1 in case of error.
* Fix _Py_RunMain(): return an exitcode rather than calling
  exit() on pymain_run_python() failure.
* _Py_ExitInitError() no longer uses ExitProcess() on Windows, use
  exit() on all platforms.
* _Py_ExitInitError() now fails with a fatal error if 'err' is not an
  error not an exit.

No NEWs is needed since the problem was introduced on master only and never released.


https://bugs.python.org/issue35589

@zooba
I just realized that this whitespace fix didn't get pushed.


https://bugs.python.org/issue36511

Similarly to how several pathlib file creation functions have an "exists_ok" parameter, we should introduce "missing_ok" that makes removal functions not raise an exception when a file or directory is already absent.  IMHO, this should cover Path.unlink and Path.rmdir.  Note, Path.resolve() has a "strict" parameter since 3.6 that does the same thing. Naming this of this new parameter tries to be consistent with the "exists_ok" parameter as that is more explicit about what it does (as opposed to "strict").


https://bugs.python.org/issue33123

Plistlib currently throws an exception when asked to decode a valid
.plist file that was generated by Apple's NSKeyedArchiver. Specifically,
this is caused by a byte 0x80 (signifying a UID) not being understood.

This fixes the problem by enabling the binary plist reader and writer
to read and write plistlib.UID objects.

Remove UNTESTED_CORE_CONFIG from test_embed.InitConfigTests: all core
config fields are now tested!

Changes:

* Test also dll_path on Windows
* Add run_main_config unit test: test config using _Py_RunMain().

test_embed: add test_init_read_set() to test newly added APIs: test
module_search_paths and executable.

Clarify how to capture stdout and stderr combined into one stream.

Python 3.8 now respects the x86-64 ABI: memory allocations are
aligned on 16 bytes. The clang flag was only used as a temporary
workaround.

Argparse can handle default value as stdin and stdout for parameters
as file type (infile, outfile).

Replace global configuration variables with core_config read from the
current interpreter.

Cleanup dynload_hpux.c.

When using multiprocessing (-jN option), worker processes now create
their temporary directory inside the temporary directory of the
main process. So the main process is able to remove temporary
directories of worker processes even if they crash or when they are
killed by regrtest on KeyboardInterrupt (CTRL+c).

Rework also how multiprocessing arguments are parsed in main.py.

This cleans the csv module a bit, I don't think it requires a bpo issue or a news entry.

Removes more legacy distutils documentation, and more clearly
marks what is left as potentially outdated, with references to
setuptools as a replacement.

Add a new pip install before `sphinx` etc.. because we should use the last version of `pip` and `setuptools`