> This is because the types of the grammars in
Lib/test/test_peg_generator/test_c_parser.py needs to be updated to use the
new sequences

I'm sorry, that doesn't mean much to me. Is it simply a matter of applying
suitable casts to the various parameters in the C code fragments in those?
For example:

*** /tmp/ediffGoXpTW 2020-09-23 08:06:38.671061988 -0500
---
/home/skip/src/python/cpython/Lib/test/test_peg_generator/test_c_parser.py
2020-09-21 15:54:43.106258797 -0500
***************
*** 325,331 ****
          compound_stmt[stmt_ty]: with_stmt
          with_stmt[stmt_ty]: (
              a='with' '(' b=','.with_item+ ')' ':' c=block {
!                 _Py_With(b, _PyPegen_singleton_seq(p, c), NULL, EXTRA) }
          )
          with_item[withitem_ty]: (
              e=NAME o=['as' t=NAME { t }] { _Py_withitem(e,
_PyPegen_set_expr_context(p, o, Store), p->arena) }
--- 325,331 ----
          compound_stmt[stmt_ty]: with_stmt
          with_stmt[stmt_ty]: (
              a='with' '(' b=','.with_item+ ')' ':' c=block {
!                 _Py_With((asdl_withitem_seq *)b, *(asdl_stmt_seq
*)*_PyPegen_singleton_seq(p,
c), NULL, EXTRA) }
          )
          with_item[withitem_ty]: (
              e=NAME o=['as' t=NAME { t }] { _Py_withitem(e,
_PyPegen_set_expr_context(p, o, Store), p->arena) }

If so, that's pretty mechanical. Even a dummy like me could produce a patch.