Don't check if the dict key exists before getting the key. Instead get the key
and handle error.

Restore the previous code for stack_overflow().

Issue #23848, #26622:

* faulthandler now only logs fatal Windows exceptions.
* write error code as decimal, not as hexadecimal
* replace "Windows exception" with "Windows fatal exception"

Only display duration if a test takes more than 30 seconds.

wcsnlen_s() result type is size_t.

Don't return a void value.

Issue #23848: On Windows, faulthandler.enable() now also installs an exception
handler to dump the traceback of all Python threads on any Windows exception,
not only on UNIX signals (SIGSEGV, SIGFPE, SIGABRT).

This function will be reused by faulthandler.

Issue #26588:

* Pass the hash table rather than the key size to hash and compare functions
* _Py_HASHTABLE_READ_KEY() and _Py_HASHTABLE_ENTRY_READ_KEY() macros now expect
  the hash table as the first parameter, rather than the key size
* tracemalloc_get_traces_fill(): use _Py_HASHTABLE_ENTRY_READ_DATA() rather
  than pointer dereference
* Remove the _Py_HASHTABLE_ENTRY_WRITE_PKEY() macro
* Move "PKEY" and "PDATA" macros inside hashtable.c

* Optimize tracemalloc_add_trace(): modify hashtable entry data (trace) if the
  memory block is already tracked, rather than trying to remove the old trace
  and then add a new trace.
* Add _Py_HASHTABLE_ENTRY_WRITE_DATA() macro

* _Py_HASHTABLE_ENTRY_DATA: change type from "char *" to "const void *"
* Add _Py_HASHTABLE_ENTRY_WRITE_PKEY() macro
* Rename _Py_HASHTABLE_ENTRY_WRITE_DATA() macro to
  _Py_HASHTABLE_ENTRY_WRITE_PDATA()
* Add _Py_HASHTABLE_ENTRY_WRITE_DATA() macro

No need to remove the old trace if the memory block didn't move.

The whole test suite takes 6 minutes on my laptop. It takes less than 30
minutes on most buildbots. The TESTTIMEOUT is the timeout for a single test
file.

libregrtest: add a watchdog to run_tests_multiprocess() using
faulthandler.dump_traceback_later().

Issue #26592: _warnings.warn_explicit() now tries to import the warnings module
(Python implementation) if the source parameter is set to be able to log the
traceback where the source was allocated.

Issue #26603:

* Implement finalizer for os.scandir() iterator
* Set the source parameter when emitting the ResourceWarning warning
* Close the iterator before emitting the warning

Issue #26604:

* Add a new optional source parameter to _warnings.warn() and warnings.warn()
* Modify asyncore, asyncio and _pyio modules to set the source parameter when
  logging a ResourceWarning warning

Add parenthesis.

Issue #26588: Enhance assertion in set_reentrant()

Issue #26588: Fix _tracemalloc start/stop: don't play with the reentrant flag.

set_reentrant(1) fails with an assertion error if tracemalloc_init() is called
first in a thread A and tracemalloc_start() is called second in a thread B. The
tracemalloc is imported in a thread A. Importing the module calls
tracemalloc_init(). tracemalloc.start() is called in a thread B.

So it's possible to get debug messages in test_tracemalloc.

Issue #26588.

Try to debug random failure on buildbots.