Fix a compilater warning on Windows 64-bit

idx variable is used for a tuple indexn so use Py_ssize_t (not int).
diff --git a/Python/ceval.c b/Python/ceval.c
index 60a71f9..e184ef6 100644
--- a/Python/ceval.c
+++ b/Python/ceval.c
@@ -2256,7 +2256,7 @@
 
         TARGET(LOAD_CLASSDEREF) {
             PyObject *name, *value, *locals = f->f_locals;
-            int idx;
+            Py_ssize_t idx;
             assert(locals);
             assert(oparg >= PyTuple_GET_SIZE(co->co_cellvars));
             idx = oparg - PyTuple_GET_SIZE(co->co_cellvars);