Issue #9548: Add a minimal "_bootlocale" module that is imported by the _io module instead of the full locale module.

time.xmlrpc.com has come and gone over the years, and has been gone again for
a while.  The test did test one thing that the current xmlrpc tests don't: the
use of multiple levels of attribute names in the call.  So in addition to
removing the network test, we add a test in xmlrpc of dotted name access.

There should also be a test for when dotted name access is disallowed, but
that requires more extensive test harness refactoring, and in any case was not
tested by the network test we are deleting, since it is a server-side setting.

This is a slightly simplified version of a patch by Vajrasky Kok.

Missed changing the doc for the command itself, patch for that also by Connor
Osborn.

So that it no longer shadows the print function.

Patch by Connor Osborn, doc and test changes by R. David Murray.

The GIL must be held to call PyMem_Malloc(), whereas PyOS_Readline() releases
the GIL to read input.

The result of the C callback PyOS_ReadlineFunctionPointer must now be a string
allocated by PyMem_RawMalloc() or PyMem_RawRealloc() (or NULL if an error
occurred), instead of a string allocated by PyMem_Malloc() or PyMem_Realloc().

Fixing this issue was required to setup a hook on PyMem_Malloc(), for example
using the tracemalloc module.

PyOS_Readline() copies the result of PyOS_ReadlineFunctionPointer() into a new
buffer allocated by PyMem_Malloc(). So the public API of PyOS_Readline() does
not change.

_PyMem_RawMalloc/Realloc/Free, instead of _PyMem_Malloc/Realloc/Free.  So it
becomes possible to use the fast pymalloc allocator for the PYMEM_DOMAIN_MEM
domain (PyMem_Malloc/Realloc/Free functions).

string. This change does nothing is most cases, but it is useful on Windows in
some cases.

Issue #18948: improve SuppressCoreFiles to include Windows crash popup suppression, and use it in more tests.
Patch by Valerie Lambert and Zachary Ware.