gh-97545: Make Semaphore run faster.

[cykerway]

gh-97545: Make Semaphore run faster.

For this to be a fifo semaphore, the asyncio loop shall run tasks in their waking order (to be confirmed).

Background:

• Issue asyncio.Semaphore waiters deque doesn't work #90155
• PR gh-90155: Fix broken asyncio.Semaphore and strengthen FIFO guarantee. #93222
• Issue Semaphore slowdown #97545
• Issue Should asyncio.{Lock,Semaphore}.locked() return True when there are waiters? #97028

Tests:

• asyncio.Semaphore waiters deque doesn't work #90155 (comment)
• asyncio.Semaphore waiters deque doesn't work #90155 (comment)
• https://gist.github.com/gvanrossum/02d9e4ef3c4f277050bf18fee0a3f189
• Lib/test/test_asyncio/test_locks.py

[gvanrossum]

I've run through this code so many times in my mind, this looks good to me. The silly speed test I wrote to flag this problem runs as fast as it used to. So I'll just merge this and backport to 3.11.

[miss-islington]

Thanks @cykerway for the PR, and @gvanrossum for merging it 🌮🎉.. I'm working now to backport this PR to: 3.11.
🐍🍒⛏🤖

[bedevere-bot]

GH-97584 is a backport of this pull request to the 3.11 branch.

[miss-islington]

Thanks @cykerway for the PR, and @gvanrossum for merging it 🌮🎉.. I'm working now to backport this PR to: 3.10.
🐍🍒⛏🤖

[bedevere-bot]

GH-97585 is a backport of this pull request to the 3.10 branch.

[kumaraditya303]

This is very vague and confusing as what Semaphore is made to run faster? threading module or asyncio?

[gvanrossum]

Sorry, that's a good point. It's asyncio.Semaphore. You should be able to create a PR that updates just that particular NEWS file. (And backport it.)