Add itertools.batched Support

[Poppro]

Adds the itertools.batched function new in python 3.12

https://docs.python.org/3/library/itertools.html#itertools.batched

[youknowone]

Thank you so much! Could you also update test_itertools from CPython 3.12 to ensure this patch is compliant its spec? The related issue is #5104

[Poppro]

@youknowone

Thank you so much! Could you also update test_itertools from CPython 3.12 to ensure this patch is compliant its spec? The related issue is #5104

I went ahead and added the 3.12 itertools_test update to this PR.

Notes:

1. This caught an out-of-order exception throw in the batched implementation. Corrected now.
2. I had to mark a number of previously passing tests as ignored, since they now expect DepreciatedWarnings to be thrown.

[youknowone]

That's a good point of tests 😄 Thank you!

[youknowone]

What happens if n is a negative number? That will be a python error rather than panic (by unwrap).
By the error type and message, you might want to change the type of BatchedNewArgs::n to usize or any unsigned type. Otherwise making an ok_or_else chain will be a good choice.

[Poppro]

Doesn't the conditional on line 1985 ensure that n will not be a number less than 1?

[youknowone]

That's right, thank you! How about n > usize::MAX?
Since unwrap is one of main source of uncontrolled panic, we prefer to unsure they don't have edge cases and a bit picky about it.
If that's logically not able to be triggered, using expect() with reasoning rather than unwrap() will be very helpful to provide the reasoning and a debug hint for future regression.

[Poppro]

Great point! CPython throws the following error in such case: OverflowError: Python int too large to convert to C ssize_t.

I've modified the unwrap to instead use ok_or and return an overflow error in this case.

[Poppro]

Build failures seem unrelated?

[fanninpm]

Build failures seem unrelated?

The macOS failure in test_re seems unrelated. However, can you please run cargo fmt to see if that fixes the other failure? (And, for good measure, consider running cargo clippy?)

[Poppro]

Build failures seem unrelated?

The macOS failure in test_re seems unrelated. However, can you please run cargo fmt to see if that fixes the other failure? (And, for good measure, consider running cargo clippy?)

Here's the results - no changes or issues generated

[youknowone]

Thank you so much!