GH-93896: always set event loop in asyncio.run and IsolatedAsyncioTestCase

[kumaraditya303]

• Issue: IsolatedAsyncioTestCase and asyncio.run no-longer call asyncio.set_event_loop #93896

[graingert]

Can you copy over the test cases from my PR?

[kumaraditya303]

Can you copy over the test cases from my PR?

I added a test.

[asvetlov]

LGTM, thanks for the fix!

[kumaraditya303]

LGTM, thanks for the fix!

Thanks for the reviews!

[graingert]

I think the docs should describe this change:

class Runner:
    """A context manager that controls event loop life cycle.
    The context manager always creates a new event loop,
    allows to run async functions inside it, sets the policy loop,
    and properly finalizes the loop at the context manager exit.
    If debug is True, the event loop will be run in debug mode.
    If loop_factory is passed, it is used for new event loop creation, however the
    policy loop will not be set

[kumaraditya303]

I think the docs should describe this change:

class Runner:
    """A context manager that controls event loop life cycle.
    The context manager always creates a new event loop,
    allows to run async functions inside it, sets the policy loop,
    and properly finalizes the loop at the context manager exit.
    If debug is True, the event loop will be run in debug mode.
    If loop_factory is passed, it is used for new event loop creation, however the
    policy loop will not be set

This would complicate things for an average user about how policy registry works. The existing docs is clear If loop_factory is passed, it is used for new event loop creation. We can iterate over docs afterwards but let's fix the release blockers first.

[miss-islington]

Thanks @kumaraditya303 for the PR, and @pablogsal for merging it 🌮🎉.. I'm working now to backport this PR to: 3.11.
🐍🍒⛏🤖 I'm not a witch! I'm not a witch!

[bedevere-bot]

GH-94608 is a backport of this pull request to the 3.11 branch.