Extend `BINARY_OP_SUBSCR_USTR_INT` to support latin1 characters

[johng]

Feature or enhancement

Currently in _BINARY_OP_SUBSCR_USTR_INT specialisation we look up just ascii tables. This results in specialisation fails when a character is latin1 encoded which might be common in many European texts.

A small extension to support latin1 would enable the specialisation to support these characters such as

PyObject *res_o = (c < 128)
                ? (PyObject*)&_Py_SINGLETON(strings).ascii[c]
                : (PyObject*)&_Py_SINGLETON(strings).latin1[c - 128];

Scanning some french text results in a 14% speedup with the test done on MacOS. This avoids any fallback to the generic path for the latin characters.

def scan(text):
    i = 0
    n = len(text)
    while i < n:
        c = text[i]          # the only str[int] site
        i += 1

ASCII  = "The quick brown fox jumps over the lazy dog. " * 20
FRENCH = "Le garçon préfère les crêpes à la fraîcheur du matin. " * 20

Has this already been discussed elsewhere?

This is a minor feature, which does not need previous discussion elsewhere

Links to previous discussion of this feature:

No response

Linked PRs

• gh-158649: Extend BINARY_OP_SUBSCR_USTR_INT to latin-1 characters #158650