Fix PyUnicode functions for Python 3.13

[cdce8p]

Replace _PyUnicode_EqualToASCIIString with PyUnicode_EqualToUTF8.

https://docs.python.org/dev/c-api/unicode.html#c.PyUnicode_EqualToUTF8
python/cpython#110289

Fixes

  /home/runner/work/mypy/mypy/mypyc/lib-rt/getargs.c: In function ‘vgetargskeywords’: (diff)
  /home/runner/work/mypy/mypy/mypyc/lib-rt/pythonsupport.h:310:45: error: implicit declaration of function ‘_PyUnicode_EqualToASCIIString’; did you mean ‘CPyUnicode_EqualToASCIIString’? [-Werror=implicit-function-declaration] (diff)
    310 | #define CPyUnicode_EqualToASCIIString(x, y) _PyUnicode_EqualToASCIIString(x, y) (diff)
        |                                             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~ (diff)
  /home/runner/work/mypy/mypy/mypyc/lib-rt/getargs.c:398:21: note: in expansion of macro ‘CPyUnicode_EqualToASCIIString’ (diff)
    398 |                 if (CPyUnicode_EqualToASCIIString(key, kwlist[i])) { (diff)
        |                     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~ (diff)

[hauntsaninja]

nit: want to change the comment on L83 of str_ops.c?

I guess we could still use _PyUnicode_FastCopyCharacters, but is maybe not the biggest deal

[cdce8p]

nit: want to change the comment on L83 of str_ops.c?

Done!

I guess we could still use _PyUnicode_FastCopyCharacters, but is maybe not the biggest deal

Yeah, technically _PyUnicode_FastCopyCharacters was only moved to internal/pycore_unicodeobject.h which we still include for _PyUnicode_EQ (see #17506). However, moving to the public PyUnicode_CopyCharacters might be the better call here (even if slightly slower if at all).
https://github.com/python/cpython/blob/v3.13.0b3/Include/internal/pycore_unicodeobject.h#L56

[cdce8p]

I guess we could still use _PyUnicode_FastCopyCharacters, but is maybe not the biggest deal

Yeah, technically _PyUnicode_FastCopyCharacters was only moved to internal/pycore_unicodeobject.h which we still include for _PyUnicode_EQ (see #17506). However, moving to the public PyUnicode_CopyCharacters might be the better call here (even if slightly slower if at all). https://github.com/python/cpython/blob/v3.13.0b3/Include/internal/pycore_unicodeobject.h#L56

Scratch that. Let's just continue using _PyUnicode_FastCopyCharacters here. It's also used in _PyUnicode_JoinArray which this functions is adapted from.