gh-131798: JIT: Narrow the return type of _BINARY_SLICE to original container type

[Zheaoli]

I'm not sure this patch will take some effect.

In my original thought, I think if we can setup the return type to the original container type(tuple, list etc.), the optimizer will specialize some code in the test.

but the test is not work on my thought. I guess I miss something important here. cc @brandtbucher@Fidget-Spinner

[brandtbucher]

@Zheaoli, I appreciate you taking the time to do this, but somebody else is already working on _BINARY_SLICE. Part of the reason I created an issue for this was so we could coordinate the work, rather than having multiple people racing on the same tasks.

Is it okay if we close this? FYI, one issue I see here is that there is no guarantee that slicing an arbitrary object will give you the same class (we need to check for some common classes instead):

>>>importmmap>>>mm=mmap.mmap(-1, 42)
>>> type(mm)
<class'mmap.mmap'>>>>type(mm[:])
<class'bytes'>

[Zheaoli]

I appreciate you taking the time to do this, but somebody else is already working on _BINARY_SLICE. Part of the reason I created an issue for this was so we could coordinate the work, rather than having multiple people racing on the same tasks.

Sorry about this！ I have not noticed this has been assigned.

[Zheaoli]

Is it okay if we close this?

Sure!

[Zheaoli]

one issue I see here is that there is no guarantee that slicing an arbitrary object will give you the same class (we need to check for some common classes instead):

Thanks about the explain. You prove one of my guess.