gh-96349: Fix minor performance regression initializing threading.Event

[dgiger42]

There was a minor performance regression making it slightly slower to initialize a threading.Event on the main branch than it was in 3.10. I think the slowdown may have been related to zero-cost exceptions.

This PR removes the slowdown (and makes it faster than it was before).

Here's a simple benchmark with pyperf:

3.10:

../cpython3.10/python -m pyperf timeit -s 'from threading import Event' 'e = Event()'
.....................
Mean +- std dev: 1.52 us +- 0.02 us

Main branch:

../cpython/python -m pyperf timeit -s 'from threading import Event' 'e = Event()'
.....................
Mean +- std dev: 1.59 us +- 0.02 us

With this PR:

../cpython/python -m pyperf timeit -s 'from threading import Event' 'e = Event()'
.....................
Mean +- std dev: 462 ns +- 10 ns

• Issue: Minor Performance Regression Initializing threading.Event #96349

[corona10]

LGTM, no reason not to merge

Mean +- std dev: [base] 1.53 us +- 0.01 us -> [pr] 493 ns +- 11 ns: 3.11x faster