[3.12] GH-106897: Add RERAISE event to sys.monitoring. (GH-107291)

[markshannon]

Manual backport of GH-107291.

@Yhg1s as release manager.

Fixes an omission in PEP 669, resulting in an imbalance between raise and unwind/handle events for exceptions.

• Issue: No sys.monitoring RAISE event emitted when exception is re-raised (3.12b4) #106897

[markshannon]

The ABI checking tool claims that the ABI has changed.
Why is the internal layout of PyInterpreterState part of the ABI?

[Yhg1s]

The ABI checking tool claims that the ABI has changed. Why is the internal layout of PyInterpreterState part of the ABI?

Because it's reachable through symbols exposed in the ABI. The ABI checking tool is not tightly constrained, and errs on the side of caution. (I'm sure we'd all be happy if you wanted to spend some time creating an accurate checking tool with no false negatives and fewer false positives :)

[markshannon]

The ABI file diff is a quite a lot larger than I would have liked, but it seems to be OK.
@Yhg1s are you OK with merging the updated ABI file?

[markshannon]

I've waded through the Doc/data/python3.12.abi diff, and the changes are as expected, plus a lot of line number changes and some reordering.