gh-100110: Specialize FOR_ITER for tuples

[Fidget-Spinner]

Same as FOR_ITER_LIST, but it accounts for 57% of our current failures.

https://github.com/faster-cpython/ideas/blob/main/stats/pystats-python-a87c46e-2022-12-04.md#specialization-attempts-3

• Issue: Specialize FOR_ITER for tuples #100110

[netlify]

✅ Deploy Preview for python-cpython-preview canceled.

Name	Link
🔨 Latest commit	e3504e7
🔍 Latest deploy log	https://app.netlify.com/sites/python-cpython-preview/deploys/6391d802c8b963000994af92

[markshannon]

I'm not sure that I entirely trust those stats, yet.
They are probably right though, and I guess we can revert this if the stats prove to be wrong.

[markshannon]

Looks good.
Have you attempted to benchmark this?

[Fidget-Spinner]

Looks good. Have you attempted to benchmark this?

No. It's just one specialisation. I doubt anything in pyperformance will change.

[markshannon]

I'm running the stats on it.
Partly to see if this PR makes the expected difference, and partly just to see if the stats workflow does what I expect.

[markshannon]

The stats show what I would expect. No specialization failures for tuples, and an increase of hit rates for FOR_ITER from about 66% to 79%.