Issue #27278. It should fix a conversion warning.

In practice, the Linux kernel doesn't return more than 32 MB per call to the
getrandom() syscall.

Patch by John Hagen.

by Jelle Zijlstra.

managed by a contextlib.ExitStack() and one of the exit stack
generators catches and raises it in a chain, do not re-raise the
original exception when exiting, let the new chained one through.
This avoids the PEP 479 bug described in issue25782.

Casting Py_ssize_t to Py_ssize_t is useless.

Issue #27278: Fix os.urandom() implementation using getrandom() on Linux.
Truncate size to INT_MAX and loop until we collected enough random bytes,
instead of casting a directly Py_ssize_t to int.

They are not used as base classes by another tests so they
can safely be removed.

sqlite3.OptimizedUnicode is an alias for str in Python 3 and
its behavior is already tested in CheckOptimizedUnicode in
Lib/sqlite3/test/factory.py.

sqlite3.OptimizedUnicode was undocumented in 0fc10a33eb4c and
probably added back with the result of a bad merge.

containing spaces.

Previously, when IDLE was started from a console or by import, a cascade
of warnings was emitted.  Patch by Serhiy Storchaka.

Replace os.popen() with subprocess.Popen. Based on patch by Victor Stinner.

If the "gcc", "cc" or "objdump" command is not available, the code was
supposed to raise an OSError exception. But there was a bug in the code. The
shell code returns the exit code 10 if the required command is missing, and the
code tries to check for the status 10. The problem is that os.popen() doesn't
return the exit code directly, but a status which should be processed by
os.WIFEXITED() and os.WEXITSTATUS(). In practice, the exception was never
raised. The OSError exception was not documented and ctypes.util.find_library()
is expected to return None if the library is not found.

Based on patch by Serhiy Storchaka.

Patch by Gareth Rees.

Update current tests that use old pattern with assertRaises
to make them more maintainable.